article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
we begin with the description of the model under consideration . given a square lattice in @xmath8 with @xmath9 vertices , we assign arrows along each edge obeying the following rule : at every vertex two arrows point in and two arrows point out .
such a rule is called the _ ice - rule _ and it only admits six possible arrow configurations , see figure [ fig1 ] , hence the name of the model : the _ six - vertex model _ or the _ model of two - dimensional ice_. on the boundary of the lattice we impose the _ domain wall boundary conditions _ ( dwbc ) , that is all arrows on the top and bottom side of the lattice are directed inside the lattice and all arrows on the left and right side point outside , see figure [ fig2 ] for a possible arrow configuration with dwbc .
configuration with dwbc , width=264,height=264 ] to each of the six possible arrow configurations @xmath10 we assign a positive real - valued boltzmann weight @xmath11 and define the partition function @xmath1 of the model as @xmath12 with @xmath13 denoting the number of vertices with arrow configuration @xmath10 . by definition , @xmath1 depends on six parameters : the weights @xmath14 .
however recalling some conservation laws ( see for instance @xcite , @xcite or @xcite ) we can reduce the number of parameters to two , namely , @xmath15 where @xmath16 and furthermore @xmath17 thus , a general weight reduces to the two parameters @xmath18 . in order to study the phase diagram ( see figure [ fig3 ] ) of the model , we introduce the parameter @xmath19 we distinguish the following three physical phase regions : the antiferroelectric phase ( af ) , @xmath20 ; the disordered phase ( d ) , @xmath21 ; and , the ferroelectric phase ( f ) , @xmath22 . in all phases
the weights are usually parameterized in the following way : for the ferroelectric phase , @xmath23 for the antiferroelectric phase , @xmath24 and for the disordered phase ,
@xmath25 here we discuss the critical line between the disordered and antiferroelectric phase , hence @xmath26 which corresponds to @xmath27 instead of using the disordered phase parameterization in the limit @xmath28 , we choose the following parameterization for the weights @xmath29 and @xmath30 : @xmath31 the above described six - vertex model with dwbc was introduced by korepin in @xcite and further studied in @xcite and @xcite .
those works lead to a remarkable determinantal formula for the partition function with dwbc , which , with weights parameterized as in , is @xmath32 where @xmath33 is the hankel determinant , @xmath34 the determinantal formula implies that @xmath33 solves the toda equation , @xmath35 as was noticed by zinn - justin @xcite , the hankel determinant @xmath33 can be expressed in terms of the partition function of a random matrix model , and then it can be reduced to orthogonal polynomials . on the critical line under consideration
, the expression in terms of orthogonal polynomials can be obtained as follows .
we write @xmath36 as the laplace transform of a continuous measure , @xmath37 and define the potential @xmath38 .
this implies for @xmath33 @xmath39 and thus for any permutation @xmath40 acting on @xmath41 @xmath42 hence after summation and identification of a vandermonde determinant @xmath43 up to the factor @xmath44 , the expression on the right is the partition function of eigenvalues in the random matrix model with the potential @xmath45 .
introduce now monic orthogonal polynomials @xmath46 with respect to the measure @xmath47 on the real line , @xmath48 then , by using the orthogonality condition , one can simplify to @xmath49 thus , for the partition function , via , we obtain the formula @xmath50 we will prove the following asymptotics of the normalizing constants @xmath51 : [ theo1 ] as @xmath3 , @xmath52 where @xmath53 and the error terms in are uniform on any compact subset of the set @xmath54 the proof of theorem [ theo1 ] is based on the riemann - hilbert approach for the potential @xmath45 with a singularity at the origin .
@xmath45 is close to the class of the freud potentials considered in the paper of kricherbauer and mclaughlin @xcite .
our approach is somewhat different from @xcite , and it gives a more detailed asymptotic formula for @xmath51 , which is needed in the asymptotics of @xmath1 .
our second main result in the present paper is the asymptotics of @xmath1 .
[ theo2 ] on the critical line between the disordered and antiferroelectric phase region with @xmath55 , as @xmath3 , @xmath56 where @xmath57 and @xmath58 is a universal constant .
the error term in is uniform on any compact subset of the set .
_ the leading term in formula for @xmath1 , after a rescaling , coincides with the one obtained in the physics literature by bogoliubov , kitaev and zvonarev @xcite .
the error term estimate of @xmath59 is much stronger than a logarithmic estimate which can be derived from @xcite .
expression for the constant @xmath7 is completely new .
the current article is a continuation of the work of the first author with vladimir fokin in @xcite and with karl liechty in @xcite , @xcite and @xcite .
they prove conjectures of paul zinn - justin in @xcite on the large @xmath0 asymptotics of @xmath1 in different phase regions : in the disordered phase with parameterization for some @xmath60 , @xmath61 with @xmath62 and an @xmath0 independent constant @xmath63 , whose value is unknown . in the antiferroelectric phase with @xmath64 where @xmath65 and @xmath66 , @xmath67 are the jacobi theta functions with the elliptic nome @xmath68 .
the constant @xmath63 is also here unknown .
furthermore , in the ferroelectric phase for any @xmath60 @xmath69 with @xmath70 finally on the critical line between the ferroelectric and disordered phases with the parameterizations @xmath71 and @xmath72 with @xmath73 @xmath74 where @xmath75\ ] ] with @xmath76 denoting the riemann zeta function and a positive , yet unknown , constant @xmath63 . to calculate the order of the phase transition on the critical line between the disordered and antiferroelectric phase regions , we would like to compare the free energy in these phase regions as we approach the critical line .
first we have to adjust the partition function in the regions .
observe that as we approach the critical line , the parameters @xmath77 , hence @xmath78 .
we rescale @xmath1 according to : @xmath79 in the disordered phase we obtain from that @xmath80 with the same @xmath7 and @xmath81 as in , and with @xmath82 here
@xmath83 are parameterized as in . in the antiferroelectric region
we obtain from that @xmath84 with the same @xmath7 and @xmath85 as in , and with @xmath86 here @xmath83 are parameterized as in .
let us remind that @xmath87 observe that as @xmath88 , @xmath89 hence @xmath90 the free energy in the antiferroelectric region can be written as @xmath91 where @xmath92 and @xmath93 observe at this point , that @xmath94 in the disordered phase region see and @xmath95 in the antiferroelectric one see are given by the same expression , but the underlying parameterizations of @xmath29 and @xmath30 are different .
introduce now the coordinates @xmath96 on the phase diagram as @xmath97 so that @xmath98 corresponds to the disordered phase region , whereas @xmath99 to the antiferroelectric one . employing the change of variables @xmath100 defined by and in the antiferroelectric phase region , and , respectively , by and in the disordered phase region
, we set @xmath101 and on the critical line ( cf .
theorem [ theo2 ] ) , @xmath102 also , we set in the ( af ) region @xmath103 we prove the following result on the order of the phase transition on the critical line , see also the work @xcite of zinn - justin . [ theo3 ] we have the following properties : 1 . for any @xmath104 , change of coordinates implies that as @xmath105 , @xmath106 2 .
the functions @xmath107 and @xmath108 can be analytically continued in @xmath109 to a neighborhood of the point @xmath110 , and in this neighborhood @xmath111 in addition , we have that @xmath112 3 . as @xmath105 , @xmath113 is exponentially small , @xmath114 that is we have an infinite order phase transition .
the setup for the remainder of the paper is as follows : we will first proof theorem [ theo1 ] within the framework of the riemann - hilbert approach to orthogonal polynomials .
we start with a rescaling of the weight and the construction of the equilibrium measure . afterwards a sequence of transformations is carried out which allow to approximate the global solution of the riemann - hilbert problem by local model functions , and enable us to solve the initial problem via iteration .
after that , theorem [ theo2 ] is a corollary to theorem [ theo1 ] , via and , with some extra arguments enabling us to prove the @xmath59 estimate of the error term .
the proof of theorem [ theo3 ] will be given at the end of the article .
let us scale the variables in as @xmath115 , so that @xmath116 now use the rescaled monic orthogonal polynomials @xmath117 with @xmath118 thus satisfying @xmath119 and simplify similarly as in @xmath120 our original task of studying the asymptotics of orthogonal polynomials with respect to the weight function @xmath121 has therefore simply been transfered to polynomials with the underlying weight @xmath122 our strategy will now focus on the determination of the large @xmath0 asymptotics of the normalizing constants @xmath123 , and then , via @xmath124 , the evaluation of @xmath33 from .
to this end introduce the normalized counting measure @xmath125 on @xmath126 @xmath127 and rewrite the integrand in as @xmath128 with @xmath129 since we are particularly interested in the behavior of @xmath33 as @xmath3 , we expect the value of @xmath130 in to be focused in a neighborhood of the global minimum of the energy functional @xmath131 , with @xmath125 reaching over @xmath132 it is well known ( for instance @xcite , @xcite ) that the energy minimization problem @xmath133.\ ] ] has a unique solution @xmath134 , called the _ equilibrium measure_. in the given situation the underlying potential @xmath135 is convex . in this case
the support of the equilibrium measure consists of one interval , @xmath136\subset\mathbb{r}.\ ] ] for our purposes the equilibrium measure will be essential in the construction of the _ @xmath137-function _ @xmath138,\ ] ] where we choose the principal branch for the logarithm .
moreover the @xmath137-function in turn determines the equilibrium measure uniquely by the following variational conditions : there exists a real constant @xmath139 , the _ lagrange multiplier _ , such that @xmath140 this characterization of the equilibrium measure leads to a riemann - hilbert problem for @xmath141 which can be solved as @xmath142 it turns out that for the given potential , @xmath135 , the @xmath137-function , its support , the density function @xmath143 , and the lagrange multiplier @xmath139 can all be evaluated explicitly , see @xcite and @xcite .
we have from , and that @xmath144 to evaluate this integral we use the residue theorem , @xmath145 where @xmath146 denotes a closed contour around the interval @xmath147 $ ] such that both points @xmath148 and @xmath149 lie to the left of @xmath146 . similarly @xmath150 and the latter integrand has an explicit antiderivative .
we conclude @xmath151\ ] ] choosing the principal branch for the square roots in . in order to evaluate the endpoints @xmath152 of the support , use the resolvent @xmath153 and compare with the expansion in as @xmath154 .
one obtains @xmath155 for the @xmath137-function itself we use an indefinite integration by parts , @xmath156 where from @xmath157 thus , @xmath158,\ ] ] and the constant @xmath159 is obtained from a comparison of the asymptotics in @xmath160 with the asymptotics in .
we move on to the density and use the resolvent equation again : @xmath161.\ ] ] however , from , @xmath162,\ ] ] since @xmath163 hence , via , @xmath164 thus the density has a logarithmic singularity at the origin . finally , for the lagrange multiplier , we state first , as a consequence of and , that for @xmath165 $ ] , @xmath166 but for such @xmath148 @xmath167 and @xmath168 hence @xmath169 we have now gathered enough information to start the asymptotical analysis .
the following riemann - hilbert problem ( rhp ) , originally introduced by fokas , its and kitaev in @xcite , is essential : find a @xmath170 piecewise analytic matrix - valued function @xmath171 such that * @xmath172 is analytic for @xmath173 * orienting the real line from left to right the following jump relation holds @xmath174 * as @xmath154 , we have @xmath175 the stated problem has a unique solution @xmath172 given by @xmath176 where @xmath177 is the @xmath178 monic orthogonal polynomial with respect to the measure @xmath179 and @xmath180 furthermore @xmath181 admits a full asymptotic expansion near infinity of the following form @xmath182 and since in our situation @xmath183 , we obtain @xmath184 which , in terms of the previous discussion in the last section , shows that we need to solve the given rhp for @xmath185 .
this solution will be obtained in the framework of the deift - zhou nonlinear steepest descent method @xcite using techniques developed in @xcite .
these techniques allow to approximate the global solution @xmath186 by local model functions , _ parametrices _ , and the iterative solution of a singular integral equation
. we will elaborate the required steps in the following subsections .
recall and make the following normalizing substitution in the orginal @xmath187-rhp @xmath188 this leads to a rhp for the function @xmath189 * @xmath189 is analytic for @xmath190 . *
the above properties of @xmath191 , see , imply the following jumps @xmath192\ ] ] and @xmath193.\ ] ] * at infinity , @xmath189 is normalized @xmath194 let us take a closer look at the listed jump matrices : for @xmath195 @xmath196 and for @xmath197 @xmath198 thus ( see ) @xmath199\ ] ] and we conclude for @xmath200 fixed , @xmath201 where the stated convergence is , in fact , exponentially fast .
secondly the line segment @xmath147 $ ] : the given potential @xmath202 admits analytical continuation in the left and right halfplane separately via @xmath203 and this continuation is however two - valued on the imaginary axis . since @xmath204 $ ] , the function @xmath205 admits analytical continuation in a neigborhood of the segment @xmath147 $ ] into the first and second quadrant of the complex plane using the appropriate continuation of @xmath206 in . on the other hand from @xmath207 and therefore @xmath208
i.e. the stated local continuation of @xmath209 into the first and second quadrant satisfies @xmath210 in the lower halfplane we argue in a similar fashion , indeed @xmath211 admits local continuation into the third and fourth quadrant satisfying @xmath212 these continuations will now be used in the following matrix factorizations @xmath213 motivating the second transformation of the rhp .
let @xmath214 denote the _ upper ( lower ) lense _ , shown in figure [ fig4 ] , which is bounded by the contour @xmath215 .
define @xmath216 then @xmath217 solves the following rhp * @xmath217 is analytic for @xmath218 , with @xmath219 and @xmath220 remains fixed . * the following jumps hold , with orientation fixed as in figure [ fig4 ] @xmath221 $ , } \smallskip\\ \bigl(\begin{smallmatrix } 1 & e^{n(g_++g_--v - l)}\\ 0 & 1\\ \end{smallmatrix}\bigr ) , & \hbox{$z\in\mathbb{r}\backslash[\alpha,\beta]$ , } \smallskip\\ \bigl(\begin{smallmatrix } 1 & 0\\ e^{ng(z ) } & 1\\ \end{smallmatrix}\bigr ) , & \hbox{$z\in\gamma_3 ^ -\cup\gamma_4 ^ -$.}\\ \end{array } \right.\ ] ] furthermore on the vertical line segment oriented upwards @xmath222 where @xmath223 with @xmath224 and similarly @xmath225 with @xmath226 where @xmath227 * at infinity , @xmath228 . with and
, the jump matrices will approach the identity matrix exponentially fast on the upper lense boundary @xmath229 and on the lower lense boundary @xmath230 .
also , as we have seen earlier in , a similar statement holds along @xmath231 $ ] .
let us direct our attention therefore towards the vertical line segment @xmath232 .
for @xmath233 we have @xmath234 where we introduced @xmath235 and with @xmath236 also @xmath237 with @xmath238 hence , combining and @xmath239 with functions @xmath240 which are analytic in a full neighborhood of the origin , satisfyting @xmath241 . as a result of an optimization consideration
, the function @xmath242 satisfies the important estimation @xmath243 hence also on the vertical line segment , the given jump matrices will , as @xmath3 , eventually approach the identity matrix .
all together we expect , and this will be justified rigorously , that as @xmath3 , @xmath217 converges to a solution of the model rhp , in which we only have to deal with the constant jump matrix on the punctured line segment @xmath244 $ ] .
let us now consider this model rhp .
find the piecewise analytic @xmath170 matrix valued function @xmath245 such that * @xmath245 is analytic for @xmath246 $ ] * along @xmath147 $ ] , the following jump holds @xmath247\ ] ] * @xmath245 has at most logarithmic singularities at the endpoints @xmath248 * @xmath249 a solution to this problem can be obtained by diagonalization @xmath250 with @xmath251 defined on @xmath252 $ ] with its branch fixed by the condition @xmath253 as @xmath154 .
fix a small neighborhood @xmath255 of the point @xmath256 and observe that @xmath257 as @xmath258 . simultaneously , @xmath259 as @xmath260 and @xmath261 where the function @xmath262 is defined for @xmath263 $ ] and fixed by the condition @xmath264 the stated local behaviors suggest to use the airy function @xmath265 in the construction of an edge point parametrix . to this end
first recall ( see for instance @xcite ) that the function @xmath265 is a solution to the airy equation @xmath266 uniquely fixed by its asymptotics as @xmath267 and @xmath268 @xmath269 as well as for @xmath270 @xmath271 now introduce for @xmath272 @xmath273 and observe from the previously stated asymptotics that as @xmath267 for @xmath274 @xmath275e^{-\frac{2}{3}\zeta^{3/2}\sigma_3}.\end{aligned}\ ] ] on the other hand if @xmath276 , we have @xmath277e^{-\frac{2}{3}\zeta^{3/2}\sigma_3}\begin{pmatrix } 1 & 1\\ 0 & 1\\ \end{pmatrix},\ \ \ \zeta\rightarrow\infty.\end{aligned}\ ] ] we assemble the following model function @xmath278 leading to the rhp depicted in figure [ fig5 ] which can be solved explicitly using airy functions , width=188,height=151 ] * @xmath279 is analytic for @xmath280 * the following jumps hold , with the contours oriented as shown in figure [ fig5 ] @xmath281 * in order to determine the behavior of @xmath279 at infinity we make the following observations .
first let @xmath282 and consider @xmath283 however here @xmath284 , hence the given product approaches the identity exponentially fast as @xmath267 . secondly for @xmath285 @xmath286 and also in this situation @xmath284 . both cases together with the previously stated asymptotics for @xmath287
imply therefore @xmath288e^{-\frac{2}{3}\zeta^{3/2}\sigma_3}.\end{aligned}\ ] ] as @xmath267 in a full neighborhood of infinity .
the model function @xmath279 will be useful in the construction of the parametrix to the solution of the @xmath289-rhp in a neighborhood of @xmath254 .
we proceed in two steps .
first define @xmath290 this change of variables is locally conformal , since @xmath291 and it enables us to define the right parametrix @xmath292 near @xmath254 by the formula : @xmath293 with @xmath294 as in and the matrix multiplier @xmath295 by construction , in particular since @xmath296 is analytic in a neighborhood of @xmath254 , the parametrix @xmath292 has jumps along the curves depicted in figure [ fig6 ] , and we can always locally match the latter curves with the jump curves of the original rhp . also these jumps are described by the same matrices as in the original rhp , indeed @xmath297 hence the ratio of @xmath217 with @xmath292 is locally analytic , i.e. @xmath298 let us explain the role of the left multiplier @xmath296 in the definition .
observe that @xmath299 this relation together with the asympotics implies that , @xmath300\nonumber\\ & = & \bigg[i+\frac{1}{96\zeta^{3/2}}\begin{pmatrix } 7\delta^{-2}-5\delta^2 & i(7\delta^{-2}+5\delta^2)\\ i(7\delta^{-2}+5\delta^2 ) & -(7\delta^{-2}-5\delta^2)\\ \end{pmatrix}+\frac{35}{4608\zeta^{6/2}}\begin{pmatrix } -1 & 12i\\ -12i & -1\\ \end{pmatrix}\nonumber\\ & & + o\big(\zeta^{-9/2}\big)\bigg]m(z)\end{aligned}\ ] ] as @xmath3 and @xmath301 ( so @xmath302 ) .
since the function @xmath294 is of order @xmath303 on the latter annulus and @xmath304 is bounded , equation yields the matching relation between the model functions @xmath292 and @xmath245 , @xmath305 which is crucial for the succesful implementation of the nonlinear steepest descent method as we shall see later on .
this is the reason for chosing the left multiplier @xmath296 in in the form .
this time the construction is similar to the one given in the last subsection . first introduce for @xmath272 @xmath307 and obtain for @xmath274 as @xmath267 @xmath308 e^{-\frac{2}{3}i\zeta^{3/2}\sigma_3}\begin{pmatrix } 1 & -1\\ 0 & 1\\ \end{pmatrix}\end{aligned}\ ] ] as well as for @xmath309 @xmath308 e^{-\frac{2}{3}i\zeta^{3/2}\sigma_3},\ \
\zeta\rightarrow\infty.\end{aligned}\ ] ] next , instead of , define @xmath310 which solves the rhp of figure [ fig7 ] which can be solved explicitly using airy functions , width=188,height=151 ] * @xmath311 is analytic for @xmath312 * we have the following jumps on the contour depicted in figure [ fig7 ] @xmath313 * a similar argument as given in the construction of @xmath279 implies @xmath314 e^{-\frac{2}{3}i\zeta^{3/2}\sigma_3},\ \
\zeta\rightarrow\infty \end{aligned}\ ] ] valid in a full neighborhood of infinity .
again we use the model function @xmath311 in the construction of the parametrix to the solution of the @xmath289-rhp near @xmath306 . instead of @xmath315
this change of the independent variable is locally conformal @xmath316 and allows us to define the left parametrix @xmath317 near @xmath306 by the formula : @xmath318 with the matrix multiplier @xmath319 similar to the situation in the last subsection , @xmath206 has jumps on the contour depicted in figure [ fig8 ] which are described by exactly the same jump matrices as in the @xmath289-rhp , hence the ratio of parametrix @xmath206 with @xmath217 is locally analytic @xmath320 and the left multiplier in provides us with the following asymptotic matchup between @xmath206 and @xmath245 @xmath321\nonumber\\ & = & \bigg[i+\frac{i}{96\zeta^{3/2}}\begin{pmatrix } 5\delta^{-2}-7\delta^2 & i(5\delta^{-2}+7\delta^2)\\ i(5\delta^{-2}+7\delta^2 ) & -(5\delta^{-2}-7\delta^2)\\ \end{pmatrix}+\frac{35}{4608\zeta^{6/2}}\begin{pmatrix } 1 & 12i\\ -12i & 1\\ \end{pmatrix}\nonumber\\ & & + o\big(\zeta^{-9/2}\big)\bigg]m(z)\end{aligned}\ ] ] as @xmath3 and @xmath322 , thus @xmath323 at this point we can use the model functions @xmath324 and @xmath206 to employ the final transformation .
in this final transformation we put @xmath325 where @xmath326 is fixed . with @xmath327 denoting the clockwise oriented circles shown in figure [ fig9 ] , the ratio - function @xmath328 solves the following rhp , width=340,height=113 ] * @xmath328 is analytic for @xmath329 with @xmath330 * for the jumps , along the infinite branches @xmath331 @xmath332 on the vertical line segment @xmath232 @xmath333 on the upper lense boundary @xmath334 respectively lower lense boundary @xmath335 @xmath336 and on the clockwise oriented circles @xmath327 @xmath337 * in a neigborhood of infinity , we have @xmath338 . here , by construction , the function @xmath328 has no jumps inside of @xmath339 and @xmath340 and across the line segment in between .
it is important to recall the previously stated behavior of the jump matrices as @xmath3 .
in fact , on the lense boundaries , the vertical line segment @xmath232 and the half rays @xmath331 the stated jump matrices approach the identity matrix . also in virtue of and the same holds true on the circles @xmath327 , together , with @xmath341 denoting the jump matrix in the latter ratio - rhp and @xmath342 the underlying contour , @xmath343 with a constant @xmath344 whose value is not important .
the latter estimation enables us to solve the ratio - rhp iteratively .
the stated rhp for the function @xmath328 * @xmath328 is analytic for @xmath345 . * along the contour depicted in figure [ fig9 ] @xmath346 * as @xmath154 , we have @xmath347 .
is equivalent to the singular integral equation @xmath348 and by standard arguments ( see @xcite ) we know that for sufficiently large @xmath0 the relevant integral operator is contracting and equation can be solved iteratively in @xmath349 .
moreover , its unique solution satisfies @xmath350 observe that for @xmath345 @xmath351 and also as @xmath3 following from , , and as well as the previous discussion about exponentially small contributions @xmath352 we are now ready to prove the statement on the large @xmath0 asymptotics of @xmath51 given in theorem @xmath353 .
we recall the following identities , see , @xmath354 with @xmath355 and trace back the transformations @xmath356 thus @xmath357 and since @xmath358 and similarly @xmath359 we have in combination with and as @xmath3 @xmath360 we start computing the contributions arising from the circles @xmath327 . from as @xmath3 @xmath361 and since @xmath362 for @xmath363 , we obtain via residue theorem @xmath364 combined together as @xmath3 @xmath365 for the integral over @xmath339 we follow the same strategy .
first from as @xmath3 @xmath366 and with @xmath367 for @xmath368 , we deduce from residue theorem @xmath369 so together @xmath370 adding and we have thus @xmath371 let us now move on to the evaluation of the integral @xmath372 as we see this evaluation requires the asymptotical computation of the integrals @xmath373 for a function @xmath374 which is analytic on @xmath375 .
consider the first integral , we have from @xmath376 where @xmath377 let us perform the change of variables @xmath378 we have @xmath379 and would now like to express @xmath109 as a function of @xmath380 . to this end
introduce @xmath381 , thus from @xmath382 and we are going to solve this equation for @xmath383 by iteration @xmath384 first @xmath385 and secondly @xmath386 this asymptotic behavior persists for subsequent @xmath387 s , we have a solution to equation of the form @xmath388 as @xmath389 .
this in turn implies for the solution @xmath390 that as @xmath391 , @xmath392,\ ] ] as well as for its derivative @xmath393.\nonumber\end{aligned}\ ] ] at this time we go back to the given integral @xmath394 and introduce @xmath395 [ prop1 ] as @xmath3 @xmath396\end{aligned}\ ] ] we use the change of variables , as well as the substitution @xmath397 , @xmath398ds\end{aligned}\ ] ] since @xmath399 can be chosen arbitrarily small , we can expand the integrand in reciprocal powers of @xmath400 .
this gives as @xmath3 @xmath401\end{aligned}\ ] ] with @xmath402 up to an exponentially small error , we have @xmath403 and @xmath404 , where @xmath405 denotes euler s constant , thus @xmath406,\ \
\ n\rightarrow\infty.\end{aligned}\ ] ] back to the integral under consideration , we notice from @xmath407 hence via proposition [ prop1 ] @xmath408 with @xmath409 the second integral in can be treated in a similar way .
indeed we have @xmath410 with @xmath411 and we deduce as @xmath3 @xmath412 with @xmath413 adding and , we end up with @xmath414 the latter expansion enables us now to evaluate .
since @xmath415 we obtain @xmath416 as @xmath3 with @xmath417 all together from @xmath418 and in particular @xmath419.\end{aligned}\ ] ] the latter expansion allows us to deduce the asymptotics of the normalization constants @xmath51 as @xmath3 . since @xmath420 one obtains @xmath421,\ \ n\rightarrow\infty\end{aligned}\ ] ] or in other words , recalling the definitions of @xmath152 and @xmath139 and the identity
@xmath422 we deduce @xmath423 \ ] ] here the stated expansion as @xmath3 is uniform on any compact subset of the set .
furthermore by stirling s approximation @xmath424 one obtains @xmath425 with @xmath426 thus proving theorem [ theo1 ] .
we go back to @xmath427 and derive in the given situation from @xmath428 with a positive , @xmath0 independent , constant @xmath7 .
applying now euler s summation formula , we have as @xmath3 @xmath429 with an @xmath0 independent term @xmath430 .
the sum can be further estimated using summation by parts : @xmath431 with @xmath432 following the notation of proposition [ prop1 ] , as @xmath3 @xmath433 hence @xmath434 and therefore the series @xmath435 is absolutely and uniformly convergent on any compact subset of the set . back to using proposition [ prop1 ] @xmath436 with an @xmath0 independent term @xmath146 ,
thus @xmath437 back to , we get as @xmath3 @xmath438 as mentioned before , the @xmath6 dependency of @xmath7 will be derived from the toda equation .
it implies @xmath439 and we now use equation @xmath440,\ n\rightarrow\infty\end{aligned}\ ] ] uniformly on any compact subset of . here
@xmath441 has already been computed , but not @xmath442 . substituting into
, we get @xmath443^{-2}\bigg[1+\frac{(-1)^n\cos\pi xn\ \cos\frac{\pi x}{2}}{n(\ln n)^2}\nonumber\\ & & \times\bigg\{1-\frac{2\ln\ln n}{\ln n}+\frac{c_0(x)}{\ln n}+o\big(\bigg(\frac{\ln\ln
n}{\ln n}\bigg)^2\big)\bigg\}\nonumber\\ & & -\frac{1}{4n^2}+o\bigg(\frac{1}{n^2(\ln n)^2}\bigg)\bigg],\ \ n\rightarrow\infty.\end{aligned}\ ] ] since @xmath444 we can simplify further @xmath445,\ \ n\rightarrow\infty.\end{aligned}\ ] ] at this point we notice that @xmath446 as well as @xmath447 which implies with @xmath448 and the error term in this identity is uniform for any @xmath6 chosen from a compact subset of .
thus from @xmath449 and by integration with respect to @xmath6 @xmath450 where @xmath451 and @xmath452 in general depend on @xmath0 , but not on @xmath6 . substituting into , we obtain that @xmath453 and @xmath7 does not depend on @xmath0 . for any @xmath454
this implies @xmath455 i.e. the limit @xmath456 exists , hence also the limit @xmath457 exists . by taking the limit @xmath3 in
, we obtain that @xmath458 thus proving [ prop2 ] the constant @xmath7 in the asymptotic formula has the form @xmath459 we will now show that , in fact , @xmath460 . to this end
recall the initial izergin - korepin formula and , in our situation @xmath461 it shows , that @xmath1 as a function of @xmath6 is even , hence in the notation of proposition [ prop2 ] , we conclude @xmath462 thus proving theorem [ theo2 ] .
we start with change of variables .
it implies in the disordered phase region , via , @xmath463 hence for @xmath98 , @xmath464 on the other hand , in the antiferroelectric phase region , via , @xmath465 hence for @xmath99 , @xmath466 the functions @xmath467 and @xmath468 are both entire , satisfying the usual relations , @xmath469 this implies that the inverse functions , @xmath470 and @xmath471 , are analytic at @xmath472 , and they satisfy the relations , @xmath473 let us make the change of variable @xmath474 in the disordered phase and @xmath475 in the antiferroelectric phase , where @xmath476 .
then implies that @xmath477 while from , @xmath478 here the both functions , @xmath479 and @xmath480 , are analytic at @xmath481 for any @xmath104 , satisfying the relations , @xmath482 now , from and we have that @xmath483 observe that for any @xmath104 , the functions @xmath484 are both analytic at @xmath481 , satisfying the relations , @xmath485 the four equations , , , , and , prove stated behavior of @xmath405 and @xmath486 as @xmath105 .
next we go back to and and employ change of variables , @xmath487 we first notice that the functions @xmath488 are both analytic at @xmath481 and @xmath489 similarly , the ratios @xmath490 in the corresponding phase regions , are analytic at @xmath481 satisfying the same relations as in , @xmath491 but this shows that the functions @xmath492 are analytic at @xmath481 and they satisfy again the same relations , @xmath493 so in a neighborhood of @xmath481 we have the following convergent taylor expansions : @xmath494 substituting back @xmath109 , and noticing that @xmath495 is clearly analytic at @xmath110 , we obtain @xmath496 with @xmath497 and the stated series in is convergent for small @xmath109 .
p. bleher and k. liechty , exact solution of the six - vertex model with domain wall - boundary conditions .
critical line between ferroelectric and disordered phases , _ j. statist .
* 134 * ( 2009 ) , 463 - 485 .
p. a. deift , t. kriecherbauer , k. t - r .
mclaughlin , s. venakides and x. zhou , uniform asymptotics for polynomials orthogonal with respect to varying exponential weights and applications to universality questions in random matrix theory , _ commun .
pure appl .
_ * 52 * ( 1999 ) , 1335 - 1425 . | in the present article we obtain the large @xmath0 asymptotics of the partition function @xmath1 of the six - vertex model with domain wall boundary conditions on the critical line between the disordered and antiferroelectric phases . using the weights @xmath2
, we prove that , as @xmath3 , @xmath4 , where @xmath5 is given by an explicit expression in @xmath6 and the @xmath6-dependency in @xmath7 is determined .
this result reproduces and improves the one given in the physics literature by bogoliubov , kitaev and zvonarev @xcite .
furthermore , we prove that the free energy exhibits an infinite order phase transition between the disordered and antiferroelectric phases .
our proofs are based on the large @xmath0 asymptotics for the underlying orthogonal polynomials which involve a non - analytical weight function , the deift - zhou nonlinear steepest descent method to the corresponding riemann - hilbert problem , and the toda equation for the tau - function . |
the trapping reaction , @xmath13 , where @xmath2 and @xmath3 are diffusing particles with diffusion constants @xmath4 and @xmath5 respectively , has seen a resurgence of interest in the last few years @xcite .
the problem is related to the two - species annihilation reaction @xcite , @xmath14 , in the case where the initial particle densities satisfy @xmath15 . in the long - time limit @xmath16 , and to study the asymptotic decay of the @xmath2-particle density it suffices to study the survival probability , @xmath6 , of a single @xmath2 particle diffusing in a sea of @xmath3 particles . in this limit , the two - species annihilation reaction is equivalent to the trapping reaction . as the initial condition , the @xmath3-particles are taken to be randomly distributed in space with uniform density @xmath10 , i.e. the probability to find a @xmath3-particle in a volume element @xmath17 is @xmath18 .
the problem of determining the asymptotic behavior of @xmath6 has only recently been solved @xcite .
the solution is limited to space dimensions @xmath19 .
the asymptotic _ form _ of the solution has been known for some time @xcite .
the @xmath2-particle survival probability is given by @xmath20 but the values of the constants @xmath21 were not known . in refs.@xcite ,
these were determined for @xmath19 by deriving upper and lower bounds on @xmath6 and showing that they pinch as @xmath22 .
the restriction to @xmath19 is a consequence of the recurrence properties of random walks . for @xmath23 ,
all particles can be treated as zero - size points . as a consequence , the combinations @xmath24 and @xmath25 are the only dimensionless quantities in the theory .
a key result of @xcite is that the coefficients @xmath21 in eq .
( [ bl ] ) are _ independent of @xmath4 _ for @xmath26 , and therefore have the form @xmath27 , where the quantities @xmath28 are pure numbers .
the result of @xcite is @xmath29 for @xmath30 , the @xmath2-particle has to be given a non - zero size , but this size only enters as a cut - off in the logarithm in eq .
( [ bl ] ) .
although these results are asymptotically exact , numerical studies @xcite show that the asymptotic regime is not reached , either in @xmath1 or @xmath30 , after any time accessible to simulation , even using sophisticated simulation methods @xcite that allow values of @xmath6 as small as @xmath31 to be achieved .
it is natural , therefore , to try to calculate the leading correction to the asymptotic behavior .
this is the principal aim of this paper .
it should be stated at the outset that this is a very challenging problem , and we have obtained concrete results only in the limit @xmath32 . the starting point of our approach is an equation , first presented in @xcite , that implicitly determines the survival probability of the @xmath2-particle _ for a given @xmath2-particle trajectory _ , averaged over all the @xmath3-particle trajectories and @xmath3-particle initial conditions . in this equation , therefore , the @xmath3-particles have been _ eliminated from the problem_. their presence is felt only through their density , @xmath10 , and their diffusion constant , @xmath5 , which enters the equation through the @xmath3-particle diffusion propagator .
the final step in the calculation is to average the survival probability over the @xmath2-particle trajectories , weighted with the wiener measure . in the present work we focus on computing the leading pre - asymptotic corrections to the known asymptotic behavior . for simplicity
we specialize to dimension @xmath1 , where the best numerical results are available . including the new correction , the result we obtain is fully consistent with the available data .
the paper is organized as follows . in section
ii we set up the necessary formalism , eliminating the @xmath3 particles from the problem . in order to make the paper self - contained , in section iii we briefly review the calculation of the leading asymptotic behavior .
the pre - asymptotic corrections are discussed in section iv , at leading nontrivial order in the ratio @xmath25 . in the short section
v we discuss how one might go beyond this leading order , and propose a general form for the pre - asymptotic correction that interpolates between small and large @xmath25 .
the paper concludes with a brief discussion and summary of the results .
the key step in deriving the ` fundamental equation ' from which we start is a conceptual one .
we treat the @xmath2 and @xmath3 particles _ as if they were non - interacting _ , and exploit the initial condition that each @xmath3 particle is randomly located anywhere in the system .
consider , for simplicity , a system of finite length , @xmath33 , containing @xmath34 @xmath3-particles ( diffusion constant @xmath5 ) , randomly distributed within it , and a single @xmath2 particle ( diffusion constant @xmath4 ) , initially located at the origin .
let @xmath35 be the @xmath2-particle s trajectory , and let @xmath36 be the probability that a given @xmath3 particle , starting at @xmath37 , has met the @xmath2 particle before time @xmath7 .
the average of this quantity over the initial position , @xmath37 , is @xmath38 , where @xmath39 is an implicit functional of @xmath35 .
the probability that @xmath40 distinct @xmath3 particles have met the @xmath2 particle , averaged over their initial positions , is @xmath41 . taking the limit @xmath42 , @xmath43 , with @xmath44 and @xmath40 held fixed , yields the poisson distribution , @xmath45 with @xmath46 = \rho r[z]$ ] , the notation emphasizing that @xmath47 is a _
functional _ of the @xmath2-particle trajectory @xmath35 .
the poisson distribution implies that @xmath46 $ ] is the mean number of distinct @xmath3 particles that have met the @xmath2-particle up to time @xmath7 .
its derivative , @xmath48 , is the crossing rate : @xmath49 is the probability that a @xmath3 particle meets the @xmath2 particle for the first time in the time interval @xmath50 .
the probability that the trajectory @xmath51 has survived , in the original _ interacting _ problem , is simply the probability that there have been no crossings : @xmath52 = \exp(-\mu[z]).\ ] ] finally , the survival probability @xmath6 is obtained by averaging @xmath53)$ ] over all possible @xmath2-particle trajectories @xmath51 with the wiener measure : @xmath54\right ] } { \int dz(t)\,\exp\left[-\frac{1}{4d_a}\int_0^t d\tau \left(\frac{dz}{d\tau}\right)^2 \right]}\ , \label{q}\ ] ] where the path - integrals are restricted to functions @xmath35 satisfying @xmath55 .
the ` action functional ' @xmath56 = \frac{1}{4d_a}\int_0^t d\tau\ , \left(\frac{dz}{d\tau}\right)^2 + \mu[z],\ ] ] appearing in the numerator of eq .
( [ q ] ) defines an effective dynamics for surviving @xmath2-particle trajectories .
the functional @xmath46 $ ] can be determined as follows .
we calculate , in two ways , the probability density to find a @xmath3 particle at the end - point of the trajectory , @xmath35 , at time @xmath7 .
first , since the particles are treated as non - interacting , and the @xmath3 particles start in a steady - state configuration of uniform density @xmath10 , this probability density is simply @xmath10 .
it follows that @xmath57 where , on the right - hand side , @xmath58 is the probability for a @xmath3 particle to have its first encounter with the @xmath2 in the time interval @xmath59 , and @xmath60^{1/2 } } \nonumber \\ & & \times \exp\left[-\frac{[z(t)-z(t')]^2}{4d_b(t - t')}\right ] \label{green}\end{aligned}\ ] ] is the probability density for this @xmath3-particle to subsequently arrive at @xmath35 at time @xmath7 .
( [ fundamental ] ) is an implicit equation for the functional @xmath46 $ ] ( noting that @xmath61 , since no @xmath3 particle can meet the @xmath2 particle in zero time ) .
the continuum version presented here was first derived in ref .
the remainder of the paper deals with some of the consequences of the ` fundamental equation ' ( [ fundamental ] ) .
before dealing with the pre - asymptotic behavior , which is the main focus of this paper , we briefly review the derivation of the leading asymptotic behavior first presented in refs.@xcite .
the argument is based on deriving upper and lower bounds for @xmath6 and showing that these bounds pinch for @xmath22 .
first consider the case where the @xmath2 particle is stationary ( i.e.@xmath62 ) , the so - called ` target annihilation problem ' , where @xmath63 for all @xmath7 .
this problem can be easily solved by a variety of methods @xcite . here
we treat it as a simple application of eq .
( [ fundamental ] ) .
we put @xmath64 in eq .
( [ green ] ) , and call the corresponding solution of ( [ fundamental ] ) @xmath65 .
it satisfies the equation @xmath66^{-1/2}\ . \label{mu_0}\ ] ] this equation is readily solved by taking the laplace transform of both sides , since the right - hand side has the form of a convolution .
solving for @xmath67 , and inverting the transform , the result is @xmath68 , as in ( [ bl ] ) , with @xmath69 and @xmath70 given by ( [ a ] ) .
it is intuitively clear that @xmath6 for the system with @xmath62 ( stationary @xmath2 particle ) provides an upper bound on @xmath6 for a system with @xmath71 .
this bound can be proved rigorously @xcite using eq .
( [ fundamental ] ) .
first we write @xmath72 in ( [ fundamental ] ) , and write , in a more compact notation , @xmath73^{1/2}}[1 - k(t_1,t_2)],\ ] ] where we have taken @xmath74 without loss of generality , and @xmath75 ^ 2/4d_b(t_1-t_2)\}. \label{k}\ ] ] eq .
( [ fundamental ] ) can then be rearranged in the form @xmath76,\ ] ] where the full @xmath77 appears on the right - hand side . taking laplace transforms of both sides , solving for @xmath78 , and inverting the transform
gives @xmath79 & = & \frac{1}{\pi}\int_0^t \frac{dt_1}{(t - t_1)^{1/2}}\nonumber \\ & & \times \int_0^{t_1}\frac{dt_2}{(t_1-t_2)^{1/2}}\,\dot{\mu}(t_2)k(t_1,t_2 ) .
\label{ubound}\end{aligned}\ ] ] eq .
( [ ubound ] ) is an implicit equation for @xmath80 because the full @xmath81 appears on the right - hand side .
note , however , that @xmath82 and @xmath83 ( because @xmath84 the mean number of different @xmath3 particles that have met the @xmath2 particle up to time @xmath7 is clearly a non - decreasing function ) . therefore @xmath85
\ge 0 $ ] for all paths @xmath35 , with equality when @xmath86 for all @xmath7 .
it follows that @xmath87 \ge \mu_0(t)$ ] for all paths @xmath88 , and @xmath89\ .
\label{upper}\ ] ] an equivalent result on the lattice , in any space dimension , was obtained in ref .
@xcite .
this rigorous upper bound for @xmath6 has the same asymptotic behavior as the rigorous lower bound derived in @xcite , proving that the asymptotic form of @xmath6 is the same as for the
targetproblem , where the @xmath2 particle is stationary .
we sketch the derivation of the lower bound in @xmath1 .
the results for other dimensions in the range @xmath19 can be found in @xcite . to derive the @xmath1 result
we can calculate the probability for a subset of histories for which the @xmath2 particle survives .
these are the histories in which ( i ) the region @xmath90 initially contains no @xmath3 particles ; ( ii ) no b - particles enter this region up to time @xmath7 ; and ( iii ) the @xmath2 particle stays within this region up to time @xmath7 .
the probabilities associated with ( i)-(iii ) are @xmath91 , @xmath92 $ ] , and ( asymptotically ) @xmath93 respectively .
the product of these probabilities gives , for any ` box size ' @xmath94 , a lower bound on @xmath6 .
optimizing this bound with respect to @xmath94 , at any given time @xmath7 , gives the best lower bound : @xmath95 .
\label{lower}\ ] ] since @xmath96 for @xmath1 , the second term in ( [ lower ] ) is subdominant for large @xmath7 , and the two bounds agree .
this proves that the target problem ( @xmath62 ) gives the exact asymptotic form of @xmath6 for any @xmath4 .
extensions of this result to general dimension @xmath19 are given in @xcite .
in the following section we discuss how one might calculate the leading pre - asymptotic correction to @xmath6 .
the starting point of the calculation is eq .
( [ q ] ) , which is an exact expression for the survival probability , @xmath6 , of an @xmath2 particle with diffusion constant @xmath4 moving in a sea of noninteracting @xmath3 particles with diffusion constant @xmath5 .
the functional @xmath46 $ ] appearing in this equation is determined implicitly by the fundamental equation ( [ fundamental ] ) .
it can be determined perturbatively as follows . writing @xmath46
= \mu_0[z ] + \mu_1[z]$ ] , where @xmath97 , we have shown that @xmath85 $ ] is given by eq .
( [ ubound ] ) where @xmath98 is given by eq .
( [ k ] ) . for small diffusion constant , @xmath4 , of the @xmath2 particle ,
the typical excursion of the end point , @xmath35 , of the @xmath2 particle trajectory will be small , and the function @xmath99 can be expanded in powers of @xmath100 : @xmath101 ^ 2}{4d_b(t_1-t_2 ) } -\frac{1}{2}\frac{[z(t_1)-z(t_2)]^4}{[4d_b(t_1-t_2)]^2 } + \cdots \label{kexp}\ ] ] this expansion is equivalent , as we shall see , to an expansion in powers of @xmath4 . replacing @xmath98 in the right - hand side of eq .
( [ ubound ] ) by the first term in the expansion ( [ kexp ] ) , and @xmath102 by @xmath103 , where @xmath104 , gives , to first order in @xmath4 , @xmath79 & = & \frac{\lambda_1}{8\pi d_b } \int_{0}^{t } \frac{dt_1}{(t - t_1)^{1/2 } } \nonumber \\ & & \times \int_{0}^{t_1}\frac{dt_2}{t_{2}^{1/2}(t_1-t_2)^{3/2 } } [ z(t_1)-z(t_2)]^2 , \nonumber \\ & & \label{eq : mu1quadrat}\end{aligned}\ ] ] which is a quadratic functional of the @xmath2-particle trajectory @xmath35 .
working at the quadratic level will be adequate to first order in @xmath4 . using the quadratic expression , eq .
( [ eq : mu1quadrat ] ) , for @xmath85 $ ] , we have the following expression for the survival probability : @xmath105)}{\int dz(t ) \exp(-s_0[z ] ) } \label{q1}\ ] ] where both path integrals are over all trajectories with initial points @xmath55 , but with the end points , @xmath35 , unconstrained .
the ` action functionals ' @xmath106 $ ] and @xmath107 $ ] are given by @xmath108 & = & s_0[z ] + \mu_1[z ] \\
s_0[z ] & = & \frac{1}{4d_a}\int_0^t d\tau\,{\dot{z}(\tau)}^2\ . \label{ss0}\end{aligned}\ ] ] let us call @xmath109 the ratio of path integrals appearing in eq.([q1 ] ) , i.e. @xmath110 .
the calculation of the leading - order contribution to the asymptotic behavior of @xmath109 can be obtained by restricting the sums - over - paths to those which start and end at @xmath64 , i.e. paths for which @xmath111 .
we can see this as follows .
the expression for @xmath109 can be rewritten in the form @xmath112 ) } { \int_{-\infty}^\infty dx \int_{z(0)=0}^{z(t)=x } dz(t ) \exp(-s_0[z ] ) } \label{q11}\ ] ] let @xmath113 and @xmath114 be the trajectories that minimize the functionals @xmath106 $ ] and @xmath107 $ ] respectively for boundary conditions @xmath55 , @xmath115 , and write @xmath116 in @xmath106 $ ] and @xmath117 in @xmath107 $ ] .
because both functionals are quadratic , we obtain immediately @xmath118\ } } { \int_{-\infty}^\infty dx \exp\{-s_0[z_{co}(x , t)]\ } } \nonumber \\ & & \times \frac{\int_{\tilde{z}(0)=0}^{\tilde{z}(t)=0 } d\tilde{z}(t ) \exp(-s[\tilde{z}])}{\int_{\tilde{z}(0)=0}^{\tilde{z}(t)=0 } d\tilde{z}(t ) \exp(-s_0[\tilde{z}])}. \label{q12}\end{aligned}\ ] ] in the first factor in ( [ q12 ] ) , the actions @xmath119 and @xmath120 are both proportional to @xmath121 . for the ` free ' action ,
@xmath107 = ( 1/4d_a)\int_0^t \dot{z}^2 d\tau$ ] , the path @xmath122 is given by @xmath123 and the corresponding action is @xmath124 = x^2/4d_at$ ] . for the action
@xmath106 $ ] one can show @xcite that @xmath125 \propto x^2/t^\phi$ ] with @xmath126 .
the integrals over @xmath37 in the numerator and denominator of the first factor in ( [ q12 ] ) therefore yield just powers of @xmath7 .
by contrast , the second factor gives the exponential of a positive power of @xmath7 .
this second factor provides the leading correction to asymptopia , and will therefore be the focus of the subsequent development . to summarize , therefore , we will investigate the asymptotic behavior of the quantity @xmath127)}{\int_{\tilde{z}(0)=0}^{\tilde{z}(t)=0 } d\tilde{z}(t ) \exp(-s_0[\tilde{z}])}. \label{q1hat}\ ] ] since the trajectories @xmath128 vanish at both @xmath129 and @xmath130 we can express them as a fourier sine series : @xmath131 we substitute the expansion ( [ eq : sinexp ] ) into eq .
( [ q1hat ] ) , where the functionals @xmath120 and @xmath119 are given by ( [ ss0 ] ) and @xmath85 $ ] by ( [ eq : mu1quadrat ] ) .
after the rescaling @xmath132 a_n$ ] we have : @xmath133 where the dimensionless coupling constant @xmath134 is given by @xmath135 and the matrix elements @xmath136 are given by @xmath137\,[\sin(n\pi x)-\sin(n\pi y ) ] .
\nonumber \\ & & \label{eq : amatrix}\end{aligned}\ ] ] we see that the time dependence in the sub - leading contribution @xmath138 to @xmath6 enters only through the coupling constant @xmath134 .
we note that the prefactor @xmath139 is small even for @xmath140 and @xmath141 , where the ` pre - asymptotic regime ' will be reached only for @xmath142 . at this timescale ,
the leading asymptotic result is @xmath143 .
this already gives strong hints as to why the asymptotic regime is so hard to reach .
a full solution of the problem , at the quadratic level , requires the spectrum of the matrix @xmath136 . at the moment
we lack such an exact solution , so instead we devise a two - step approach to the problem : ( i ) we obtain an analytical lower bound for @xmath138 ; ( ii ) we calculate numerically the spectrum of the matrix @xmath2 , truncated at a finite dimension @xmath144 , and we make a finite - size scaling analysis for the sub - leading correction @xmath145 at finite @xmath144 .
consider eq .
( [ eq : psquadratic ] ) . by extracting the diagonal terms ,
@xmath146 , from the quadratic form in the exponential , this equation can be rewritten as @xmath147\right\rangle\ ] ] where the average is performed with the gaussian weight @xmath148.\ ] ] here we are using the diagonal part of the action , in the fourier basis , as a trial action functional in a variational calculation .
the variational property follows from the convexity inequality , @xmath149 , which immediately implies @xmath150 where @xmath151 \label{qdiag}\ ] ] is a lower bound for @xmath138 . in the appendix we show that , for large @xmath40 , @xmath152 . for large @xmath134 ,
corresponding to large @xmath7 , the sum on @xmath40 is dominated by large values of @xmath40 , of order @xmath153 , and the sum can be replaced by an integral to leading order : @xmath154 \nonumber \\ & \approx & \exp(-ag^{2/3})\ , \label{33}\end{aligned}\ ] ] where @xmath155 using eq.([g ] ) for @xmath134 , the final result takes the form @xmath156\ . \label{newlower}\ ] ] it is interesting to compare this result with the rigorous lower bound ( [ lower ] ) .
the latter is valid for all @xmath4 , while the new bound is valid , in principle , only for @xmath157 . in both expressions
the leading correction to asymptopia is a term of order @xmath158 in the exponential , where @xmath159 has the dimensions of a diffusion constant . in the bound ( [ lower ] ) @xmath160 , while in the new bound ( [ newlower ] ) @xmath161 . for small enough @xmath25 , therefore , the new bound is tighter than the old , while for large @xmath25 the old bound is tighter .
this confirms that the new bound can not be valid in general , but only for small enough @xmath25 .
the two bounds cross when @xmath162 , so the new bound is better for the case @xmath163 used in simulations .
although the new bound is not strictly valid for @xmath164 since the condition @xmath32 is not satisfied , the above considerations suggest that @xmath165 may be sufficient in practice .
data for @xmath6 were obtained using the algorithm of mehra and grassberger @xcite .
the system parameters were @xmath166 and @xmath167 .
we plot , in figure 1 , @xmath168 against @xmath169 .
the asymptotic value of this quantity is @xmath170 .
the upper and lower bounds ( [ upper ] ) and ( [ lower ] ) are represented by the lower and upper dashed lines respectively ( note the apparent interchange of bounds due to the minus sign in the definition of the ordinate ) .
the new lower bound ( [ newlower ] ) is represented by the dotted line .
it is clearly a significant improvement over the old bound , and appears to be converging towards the data at late times . in the following subsection we provide a finite - size scaling analysis which strongly suggests that the lower bound , @xmath171 , gives the correct asymptotic form , @xmath172 , although the constant @xmath173 may differ from ( i.e. be smaller than ) that appearing in eq .
( [ newlower ] ) .
one can show that the matrix elements @xmath136 decay as a power law along the rows , so it is not obvious whether or not the off - diagonal matrix elements make a significant contribution to the eigenvalues .
these off - diagonal terms do not enter the calculation of the lower bound @xmath174 , so it is clearly of interest to know whether they qualitatively change the result . as an analytical insight is missing so far , we study this problem numerically using the eigenvalues of finite matrices @xmath175 to compute @xmath176 of eq .
( [ eq : psquadratic ] ) .
the latter equation gives @xmath177^{-1/2 } \nonumber \\ & = & \exp\left(-\frac{1}{2}\sum_{n=0}^{\infty } \ln(1+g\lambda_n)\right)\ , \end{aligned}\ ] ] where @xmath178 is the unit matrix and @xmath179 is the @xmath40th eigenvalue of @xmath2 ( where the eigenvalues are listed in decreasing order ) .
ideally , in a finite-@xmath144 analysis we would simply truncate the sum at @xmath180 . in practice , of course
, we do not know the eigenvalues explicitly , so instead we truncate the matrix @xmath2 to an @xmath181 matrix and compute the eigenvalues of the truncated matrix . in the asymptotic regime
we anticipate the finite - size scaling form @xmath182=n^{\alpha}f(g / n^\beta)\ .
\label{fss1}\ ] ] the condition that @xmath176 has an @xmath144-independent large-@xmath144 limit implies that @xmath183 for @xmath184 . it is instructive to first consider the diagonal approximation to the matrix @xmath2 , in which the off - diagonal terms are neglected .
this provides the lower bound @xmath174 given by eq .
( [ qdiag ] ) . the corresponding finite - size quantity , @xmath185 is given by @xmath186 , leading to @xmath187 = \frac{1}{2 } \sum_{n=1}^n \ln(1+ga_{nn})\ .\ ] ] for large @xmath40
, we show in the appendix that @xmath188 has the asymptotic form @xmath189 , so the finite - size scaling variable is @xmath190 , i.e. @xmath191 in ( [ fss1 ] ) within the diagonal approximation . in the limit @xmath192 , @xmath42 with @xmath190
held fixed , one can replace the sum over @xmath40 by an integral and one readily recovers the scaling form ( [ fss1 ] ) with @xmath193 , @xmath191 .
it is also easy to show that the scaling function @xmath194 in eq .
( [ fss1 ] ) has , within the diagonal approximation , the limiting forms @xmath195 for @xmath196 , as expected , while @xmath197 for @xmath198 .
our aim is to investigate whether the same general scaling structure ( [ fss1 ] ) , with the same exponents @xmath193 and @xmath191 , holds when the full matrix @xmath2 is used instead of just its diagonal part .
if the matrix @xmath2 is simply truncated to an @xmath199 matrix , however , and the eigenvalues determined , the corrections to finite - size scaling are found to be large . we attribute this to a large influence of the missing matrix elements @xmath136 , with @xmath200 or @xmath40 greater than @xmath144 , on the eigenvalues @xmath179 when @xmath40 gets close to @xmath144 . instead ,
therefore , we adopt a more refined approach in which the matrix is truncated to size @xmath201 , but we use only the first @xmath144 eigenvalues to compute @xmath202 .
furthermore , instead of computing @xmath203 , we compute the derivative @xmath204 since we find that corrections to finite-@xmath144 scaling are smaller for this quantity . the finite - size scaling form for @xmath205 follows from its definition and eq .
( [ fss1 ] ) : @xmath206 where @xmath207 . on the basis of the diagonal approximation we expect @xmath208 , @xmath209 , and @xmath210 for @xmath196 , where @xmath2 is given by eq .
( [ a ] ) , and , from eq .
( [ h ] ) , @xmath211 for @xmath212 .
we therefore plot @xmath213 against the scaling variable @xmath214 .
[ fig : psscaling ] shows that a good scaling collapse is obtained with @xmath193 and @xmath191 .
this result implies that @xmath215 , i.e. the nondiagonal elements of the matrix @xmath2 do not qualitatively change the scaling of the pre - asymptotic correction to the survival probability .
the straight lines in the figure are the asymptotes dicussed above , with slopes @xmath216 and @xmath217 for small and large scaling variable respectively .
the main curve actually contains both the full matrix ( continuous lines ) and the diagonal approximation ( symbols ) , which lie almost on top of each other .
a very small deviation from perfect scaling is evident in the data .
in fact , a slightly better fit is obtained with @xmath218 instead of 1.5 .
however , the same numerical scaling analysis applied to the diagonal matrix , where @xmath193 and @xmath191 are exact , also yields , for the range of @xmath144 and @xmath134 covered by fig.([fig : psscaling ] ) , an apparent best fit with @xmath219 .
we conclude that , for the full matrix , @xmath191 is consistent with the data within the numerical precision .
we conclude , finally , that the form of the leading correction to asymptopia given by the lower bound ( [ 33 ] ) is correct .
even the amplitude seems to be given rather accurately by the lower bound .
in the previous sections we have used the functional @xmath85 $ ] correct to first order in the ratio @xmath25 of the diffusion constant of the particle to that of the traps . to this leading order , the parameter which controls the asymptotic behaviour is @xmath220 .
we obtained the following asymptotic lower bound for the survival probability , @xmath6 , of the @xmath2-particle in @xmath1 : @xmath221\ ] ] where @xmath222 and @xmath2 is a constant .
this suggests , as discussed in the preceding section , that at large @xmath25 the new lower bound can cross the lower bound deduced from an `` excluded volume '' kind of argument @xcite , given by eq.([lower ] ) . to escape from this apparent contradiction
, one has to consider higher - order ( in @xmath25 ) terms in the action . to go to next order
, we start from eq .
( [ ubound ] ) , and expand the right - hand side up to order @xmath223 . after a tedious but straightforward calculation
, we obtain a result for the term order @xmath223 in the form of a sum of ( multiple ) integrals . inserting the fourier expansion ( [ fourier ] ) and rescaling the expansion coefficients @xmath224 as before to make them dimensionless
, we can cast the action in the dimensionless form @xmath225 ^ 2 ) , \label{s2}\ ] ] where @xmath226 and the vertices @xmath136 , @xmath227 are pure numbers .
eq.([s2 ] ) enables us to extend the calculation of @xmath138 , eq .
( [ eq : psquadratic ] ) , to one higher order in @xmath25 by replacing the exponential in eq .
( [ eq : psquadratic ] ) by @xmath228 with @xmath119 given by eq .
( [ s2 ] ) .
while we have not carried out this program explicitly , the structure of the expansion ( [ s2 ] ) shows that the quadratic approximation is valid for small @xmath25 .
what is the effect of the higher order terms ?
for the case @xmath229 ( where the traps are immobile ) the asymptotic behavior has , in @xmath1 , the known functional form @xmath230 , where @xmath231 .
it is plausible to assume that higher - order corrections change the coefficient @xmath173 but leave the exponent unaltered .
then , if the above series in @xmath25 is convergent , we obtain the asymptotic behavior of @xmath6 , including the leading pre - asymptotic correction , in the form @xmath232\ .
\label{conj}\ ] ] the function @xmath233 has the limiting behavior @xmath234 for @xmath235 , with @xmath236 from the lower bound ( [ newlower ] ) , while @xmath237 for @xmath238 , where @xmath239 to match the known asymptotic behavior with immobile traps the so - called donsker - varadhan problem @xcite .
this latter limit agrees with eq .
( [ lower ] ) ( where @xmath240 for immobile traps ) .
the reason is simple . in @xmath1 ,
only the two immobile traps immediately to the left and right respectively of the @xmath2 particle play any role . let the distances of these traps from the @xmath2-particle be @xmath37 and @xmath241 at @xmath242 .
then the particle survives as long as it stays within the box of size @xmath243 defined by these two traps , so its asymptotic survival probability , averaged over @xmath37 and @xmath241 , is given by @xmath244 $ ] . evaluating the integral by steepest descents for @xmath245
one recovers eq.([lower ] ) ( with @xmath240 ) .
in this paper we have addressed the issue of the leading correction to asymptopia in the long - time dynamics of the trapping reaction , @xmath246 .
the main results are a new rigorous lower bound , valid in the limit @xmath32 , on the survival probability of the @xmath2-particle , and a numerical analysis , based on finite - size scaling , indicating that the new lower bound contains the correct analytic form for the correction to asymptopia , namely a @xmath247 correction to the leading @xmath248 term in the exponential .
the new bound agrees well , at late times , with numerical data , as shown in figure 1 .
we also presented a conjecture , eq .
( [ conj ] ) , for the form of the correction to asymptopia for general @xmath25 , incorporating both extreme limits , @xmath249 and @xmath250 .
open questions to be addressed in future work include the generalization of the work presented here to higher spatial dimensions .
another interesting question , addressed briefly in @xcite , is the scaling with time of the fluctuations of surviving trajectories .
these can be studied starting from eq .
( [ eq : mu1quadrat ] ) , using the same fourier decomposition of the trajectory as employed in this work .
one finds that the fluctuations are subdiffusive as suggested in @xcite and observed in numerical simulations @xcite .
details of these calculations will be presented in a separate publication .
we thank richard blythe for discussions .
la acknowledges support from the european community marie curie fellowship scheme under contract no .
hpmf - ct-2002 - 01910 .
the matrix element @xmath136 is given by the double integral @xmath251 which by variable substitutions @xmath252 , @xmath253 and standard manipulations of the trigonometric functions can be reduced to a sum of one - dimensional integrals @xmath254 where @xmath255 is the bessel function of the first kind .
the fact that @xmath256 if @xmath257 can be obtained from the initial formula , eq .
( [ eq : amninit ] ) , if one notices that the kernel is symmetric at reflection about the line @xmath258 . for the diagonal elements , @xmath259 , the leading large-@xmath40 contribution comes from the first integral @xmath260 correct to leading order for large @xmath40 .
it follows that @xmath261 for @xmath262 .
m. d. donsker and s. r. s. varadhan , comm .
* 28 * , 525 ( 1975 ) ; b. d. hughes , _ random walks and random environments _ , vol . 1 ( oxford university press , new york , 1995 ) , and
references therein . | we consider the trapping reaction @xmath0 in space dimension @xmath1 , where the @xmath2 and @xmath3 particles have diffusion constants @xmath4 , @xmath5 respectively .
we calculate the probability , @xmath6 , that a given @xmath2 particle has not yet reacted at time @xmath7 . exploiting a recent formulation in which the @xmath3 particles are eliminated from the problem we find , for @xmath8 , @xmath9 $ ] , where @xmath10 is the density of @xmath3 particles and @xmath11 for @xmath12 . |
finding low energy counterparts to @xmath0-ray sources is difficult because they have large positional uncertainties , typically @xmath9 across .
consequently , the majority of the high energy @xmath0-ray sources observed by _
cos b _ and _ egret _ remain unidentified @xcite .
young pulsars are viable candidates for many of these sources , and they remain the only discrete galactic source class ( other than the sun ) unambiguously shown to emit radiation in the 100 mev
10 gev range .
the increased sensitivity to pulsars with high dispersion measure ( dm ) , provided by new pulsar backends such as the parkes multibeam system @xcite , has led to a number of discoveries recently of young pulsars coincident with known @xmath0-ray sources @xcite .
the recent detection of a young , low luminosity , radio pulsar in the supernova remnant 3c58 @xcite suggests many more faint radio pulsars await discovery in deep , targeted , searches .
the _ cos b _ source 2cg 075 + 00 has long been considered a pulsar candidate due to its hard spectrum .
_ egret _ resolved it into two sources .
the fainter one , 3eg j2016 + 3657 , was identified as a probable blazar by @xcite .
a _ rosat _ image and an _ asca _ gis image based on the second _ egret _ catalog position of the brighter source , 3eg j2021 + 3716 , failed to yield any obvious candidates .
these sources are in the crowded cygnus region and the likelihood analysis used to find @xmath0-ray source positions is sensitive to the assumed number of surrounding sources .
@xcite ( hereafter , rrk ) noted that the catalog of sources above 1 gev @xcite contained two sources in this region that were not in the list of 3 @xmath10 sources , based on 100 mev and above maps , used to construct the 3rd _ egret _ catalog .
they rederived the positional contours of 3eg j2021 + 3716 using 1 gev and above maps and all of the known nearby sources in their analysis .
an _ asca _
image of the new position yielded two hard point - like sources ( figure [ fig : xray ] ) , embedded in softer extended emission .
one is identified with the wolf - rayet + o - star binary system wr 141 . the second , ax j2021.1 + 3651 , is moderately absorbed ( n@xmath11 = @xmath12 ) with a hard , power - law spectrum ( photon index @xmath13 ) . by targetting the x - ray source ax j2021.1 + 3651 with the arecibo radio telescope
, we have discovered the young and energetic pulsar psr j2021 + 3651 @xcite .
here we present a timing solution of the pulsar based on eight months of monitoring with arecibo , and discuss the likely prospect that this is the counterpart to 2cg 075 + 00 , which has eluded identification for over twenty years .
psr j2021 + 3651 was observed 17 times between mjd 52305 and 52545 using the wideband arecibo pulsar processor ( wapp ) at the arecibo observatory .
the wapp is a fast - dump digital correlator with adjustable bandwidth ( 50 or 100 mhz ) and variable numbers of lags and sample times ( for details see dowd , sisk , and hagen 2000 ) .
our observations were made at 1.4 ghz with 100 mhz of bandwidth , 512 lags , 200 @xmath14s sampling , and summed polarizations .
the 16-bit samples were written to a disk array and then transfered to magnetic tape for later analysis .
analysis of the observations was done using the presto software suite @xciteransom/ ] .
integrated pulse profiles ( figure [ fig : pulse ] ) from our observations were convolved with a template profile to extract 69 topocentric times of arrival ( toa ) . using tempo
, the topocentric toas were converted to toas at the solar system barycenter at infinite frequency and fit simultaneously for pulsar period , period derivative , dm , ra , and dec with a residual rms of 292 @xmath14s .
table [ tab : pulsar ] gives the measured and derived parameters for psr j2021 + 3651 , and is updated from @xcite .
the quoted errors are three times those given by tempo in order to compensate for systematic errors otherwise unacounted for .
l @ c + + parameter & value + + + right ascension @xmath15 ( j2000 ) & 20@xmath16 21@[email protected](51 ) + declination @xmath19 ( j2000 ) & + 36@xmath20 51@xmath21 [email protected](72 ) + galactic longitude @xmath23 & 75.@xmath2423 + galactic latitude @xmath25 & + 0.@xmath2411 + pulse period @xmath26 ( s ) & 0.103722225646(12 ) + period derivative @xmath27 & @xmath28 + pulse frequency @xmath29 ( s@xmath5 ) & 9.6411351933(11 ) + frequency derivative @xmath30 ( s@xmath31 ) & @xmath32 + rms residual & 292@xmath14s + epoch ( mjd ) & 52407.389 + dispersion measure dm ( pc @xmath7 ) & 369.12(58 ) + pulse width at 50@xmath33 of peak w@xmath34 ( ms ) & 9.9 + pulse width at 10@xmath33 of peak w@xmath35 ( ms ) & 18 + flux density at 1425 mhz ( mjy ) & @xmath36 + spin - down luminosity @xmath37@xmath38 ( erg s@xmath5 ) & @xmath39 + surface dipole magnetic field @xmath40@xmath41 ( g ) & @xmath42 + characteristic age @xmath43 ( kyr ) & 17 + + + @xmath38 @xmath44 with @xmath45 g @xmath46 + @xmath41 assuming standard magnetic dipole spindown : @xmath47 + gauss @xcite
the dm of psr j2021 + 3651 is by far the highest known in the galactic longitude range @xmath48 , which is mainly an inter - spiral arm direction .
using the recent @xcite update to the @xcite dispersion measure model gives a distance of @xmath49 kpc , at the outer edge of the last spiral arm used in the model .
it is possible that there are further contributions from clouds in the cygnus region not included in the model , where there is known to be excess gas at @xmath50 kpc ( j. cordes , private communication , 2002 ) .
however , there are no obvious hii regions within the arecibo beam seen in either very large array ( vla ) 20-cm radio or midcourse space experiment ( msx ) 8.3 @xmath14 m images ( available from the nasa / ipac infrared science archive ) .
the likelihood that psr j2021 + 3651 is unrelated to the x - ray source is extremely small given the rarity of such young , energetic , pulsars and the consistency of the timing position with the @xmath51 positional uncertainty of the _ asca _ source .
however , the high dm is somewhat surprising given the x - ray absorption quoted by rrk ( n@xmath11=(5.0@xmath52 cm@xmath31 ) , where the errors represent the 90% confidence region .
the total galactic hi column density in this direction as estimated from the ftool nh , which uses the hi map of @xcite , is @xmath54 cm@xmath31 .
this should be a good approximation if the source is truly at the far edge of the outer spiral arm .
noting that the _ asca _ image shows faint , softer emission in the region ( figure [ fig : xray ] ) , and given the likely possibility of either associated thermal x - ray flux from a supernova remnant or a nearby massive star , we fit the _ asca _ spectrum of rrk , adding a thermal component to the absorbed power - law model .
accounting for @xmath55% of the photon flux with a mekal thermal plasma model ( see @xcite and references therein ) of temperature @xmath56 kev in xspec @xcite statistically improves the fit ( _ f_-test chance probability of 2.5% ) .
the best - fit absorption for this three component model is n@xmath11=7.6@xmath57 cm@xmath31 with a 90% confidence region of ( 4.1 12.3)@xmath57 cm@xmath31 , consistent with the total galactic column density .
the best - fit photon index is @xmath58 , still consistent with the 1.47 2.01 range in rrk derived from the simple absorbed power - law model .
hence the x - ray absorption does not force us to adopt a smaller distance than is suggested by the dm and the value of n@xmath11 may be contaminated by thermal emission .
observations of the source may well clarify this . for a distance @xmath59 kpc , the inferred isotropic x - ray luminosity @xmath60 erg s@xmath5 ( 2 10 kev ) .
the x - ray efficiency @xmath61 is 0.01@xmath62 .
compared to the total pulsar plus nebula x - ray luminosity of other spin - powered pulsars , this is somewhat high , but within the observed scatter @xcite .
upcoming chandra observations will determine what fraction of the x - ray flux comes from a compact nebula , and will allow us to search for x - ray pulsations .
2cg 075 + 00 = 3eg j2021 + 3716 = gev j2020 + 3658 is the 10th brightest galactic source above 1 gev .
it has a hard ( photon index @xmath63 ) spectrum with no sign of a break out to 10 gev , and has low variability , similar to the known @xmath0-ray pulsars .
psr j2021 + 3651 s position is outside the 99% contour of the 3eg catalog position , but consistent with the 95% contour of both the @xcite gev catalog position and the rrk position .
the only other potential counterpart within any of the error contours is wr 141 .
although such binary systems have the potential for producing @xmath0-rays from colliding winds , no convincing association with a known high - energy ( @xmath64 100 mev ) @xmath0-ray source has yet been made for a wolf - rayet system @xcite . the pulsar s positional coincidence coupled with the high inferred spin - down
luminosity strongly suggests this pulsar emits @xmath0-rays .
unfortunately , confirming this by folding archival _ egret _ data is problematic due to the likelihood of significant past timing noise and glitches , which make the back - extrapolation of the rotational ephemeris uncertain . assuming the distances given by the @xcite model , the known @xmath0-ray pulsars have @xmath0-ray efficiencies @xmath65 mostly between 0.0001 and 0.03 ( assuming 1 sr beaming ) .
assuming the @xmath0-rays are 100% pulsed , the inferred @xmath0-ray efficiency for psr j2021 + 3651 , @xmath66 in the 100 mev to 10 gev range , would be by far the most efficient @xmath0-ray pulsar .
note however that psr b1055@xmath6752 @xcite has a similar efficiency if one uses the dm distance given by @xcite , instead of the newer @xcite model .
psr b1706@xmath6744 , which has nearly identical spin parameters , has a @xmath0-ray efficiency of @xmath68 , while vela with a similar spin - down energy has a low efficiency of @xmath69 .
even if the pulsar is located within the perseus arm at a distance of 5 kpc , it would still be very efficient .
if it is the @xmath0-ray source , it would further strain the relationship of efficiency to spin - down energy @xmath70 expected from theory @xcite . while there is currently no observational evidence for a closer distance ,
increased dm from an intervening source in this relatively crowded direction would not be surprising .
we note that the dm derived distances for several other young pulsars recently discovered within _ egret _ error boxes , if the true counterparts , also tend to have high inferred @xmath0-ray efficiencies @xcite .
we plan to continue timing observations of this source , which will refine the position , provide a contemporaneous timing solution for future high - energy pulse searches , and allow us to monitor for glitches .
_ agile _ and _ glast _ should be able to unambiguously determine if the pulsar is the counterpart of the @xmath0-ray source .
we thank jim cordes for useful discussions . we acknowledge support from nserc , cfi , an nsf career award , and a sloan fellowship . m.s.e.r .
is a quebec merit fellow .
s.m.r . is a tomlinson fellow .
is an nserc pgs a fellow .
v.m.k . is a canada research chair .
the arecibo observatory is part of the national astronomy and ionosphere center , which is operated by cornell university under a cooperative agreement with the national science foundation .
camilo , f. , stairs , i. h. , lorimer , d. r. et al . , discovery of radio pulsations from the x - ray pulsar j0205 + 6449 in supernova remnant 3c 58 with the green bank telescope , _ astrophys .
j. l. _ , * 571 * , l41-l44 , 2002 .
halpern , j. p. , camilo , f. , gotthelf , e. v. et al . , psr j2229 + 6114 : discovery of an energetic young pulsar in the error box of the egret source 3eg j2227 + 6122 , _ astrophys .
j. l. _ , * 552 * , l125-l128 , 2001 .
manchester , r. n. , lyne , a. g. , camilo , f. et al . ,
the parkes multi - beam pulsar survey - i. observing and data analysis systems , discovery and timing of 100 pulsars , _ mon . not .
r. astron . soc .
_ , * 328 * , 17 - 35 , 2001 . | the _ cos b _ high energy @xmath0-ray source 2cg 075 + 00 , also known as gev j2020 + 3658 or 3eg j2021 + 3716 , has avoided identification with a low energy counterpart for over twenty years .
we present a likely identification with the discovery and subsequent timing of a young and energetic pulsar , psr j2021 + 3651 , with the wideband arecibo pulsar processor at the arecibo observatory .
psr j2021 + 3651 has a rotation period @xmath1 ms and @xmath2 , implying a characteristic age @xmath3 kyr and a spin - down luminosity @xmath4 erg s@xmath5 .
the pulsar is also coincident with the _ asca _ source ax j2021.1 + 3651 .
the implied luminosity of the associated x - ray source suggests the x - ray emission is dominated by a pulsar wind nebula unresolved by _
asca_. the pulsar s unexpectedly high dispersion measure ( dm @xmath6 pc @xmath7 ) and the @xmath8 kpc dm distance pose a new question : is psr j2021 + 3651 an extremely efficient @xmath0-ray pulsar at the edge of the galaxy ? this is a question for _ agile _ and _ glast _ to answer . |
let @xmath8 . by @xmath9
we denote the standard partial order on @xmath2 , that is , for @xmath10 one has @xmath11 if and only if for each @xmath12 the @xmath13-th component of @xmath14 is not larger than the @xmath13-th component of @xmath15 .
the notations @xmath16 and @xmath17 stand for the convex hull and interior , respectively . in
what follows @xmath1 stands for an arbitrary @xmath18-dimensional closed convex set in @xmath2 , @xmath5 for a nonempty class of @xmath18-dimensional closed convex sets in @xmath2 and @xmath0 for an arbitrary closed convex set in @xmath2 ( not necessarily @xmath18-dimensional ) . in this manuscript
we study the functionals @xmath4 , @xmath6 and @xmath19 ( with @xmath20 ) defined by @xmath21 we call @xmath3 the _ @xmath1-reduction _ , @xmath22 the _ @xmath5-closure _ and @xmath23 the @xmath13-th _ @xmath5-closure _ of @xmath0 .
the term closure in the given context goes back to chvtal @xcite .
we remark that @xmath6 can be used to define the well - known _ chvtal closure _ and _ split closure _ ( see @xcite , @xcite , ( * ? ? ?
* chapter 23 ) ) .
furthermore , @xmath4 and @xmath6 provide a natural link to _ disjunctive programming _ ( see also @xcite ) . if @xmath0 is a polytope , then @xmath3 is also a polytope , which can be obtained from @xmath0 by ` cutting off ' parts of @xmath0 with hyperplanes which are determined by @xmath1 .
see also fig .
[ r_l : illustration ] for an illustration .
let us discuss the relation of the introduced functionals to the cutting - plane theory .
we shall not discuss computational aspects but rather geometric ideas of cutting - plane methods .
we call a subset @xmath24 of @xmath2 a _ mixed - integer space _ if @xmath25 for integers @xmath26 and @xmath27 satisfying @xmath28
. a _ mixed - integer linear problem _ is an optimization problem having the following form : @xmath29 where @xmath30 , @xmath31 is a rational matrix and @xmath32 is a rational vector . the sizes @xmath31 and @xmath32 are assumed to be chosen properly so that the expression @xmath33 makes sense . in geometric terms , we are given a rational polyhedron @xmath34 and we maximize a linear function on @xmath35 . in the case @xmath36 all vertices of @xmath37 belong to @xmath24 .
thus , in this case problem can be essentially reduced to a problem of linear programming by relaxing the condition @xmath38 to @xmath39 . in the case
@xmath40 there exist closed halfspaces @xmath41 such such that @xmath42 and @xmath43 .
we call a halfspace @xmath41 as above a _ cut _ for @xmath37 and we call the boundary of @xmath41 a _ cutting plane_. if we replace @xmath37 by @xmath44 we pass to an equivalent mixed - integer linear problem ( in order to preserve the rationality of @xmath37 , we need to assume that @xmath41 is a rational polyhedron ) . in analytic terms , the system of linear inequalities @xmath45 is modified by adding a constraint which defines the cut @xmath41 . the above reduction step for @xmath37 helps ` get closer ' to the case @xmath36 since this step decreases the difference between @xmath37 and @xmath46 .
the reduction of @xmath37 to @xmath44 is the basic computational step of the _ cutting - plane methods _ ( see also @xcite for further information ) .
it turns out that , for deriving cuts , functionals @xmath4 can be used .
we call a @xmath18-dimensional closed convex set @xmath1 in @xmath2 an _ @xmath24-free set _ if @xmath47 . for an @xmath24-free set @xmath1
one obviously has @xmath48 .
thus , a halfspace @xmath41 satisfying @xmath49 and @xmath43 is a cut for @xmath37 ( see fig .
[ fig - closure - op ] ) .
an @xmath24-free set @xmath1 in @xmath2 is said to be a _
maximal @xmath24-free set _ if there exists no @xmath24-free set @xmath50 with @xmath51 . among all cuts determined by @xmath24-free sets
the strongest ones arise from maximal @xmath24-free sets .
thus , maximal @xmath24-free sets are of particular importance for the cutting - plane theory ( see @xcite for related results and fig .
[ ex - m - free ] for an illustration ) .
c = 1.1 mm ( 100,35 ) ( 0,-3 ) ( 24,28)@xmath1 ( -3,19)@xmath0 ( 65,17)@xmath3 + c ( 90,35 ) ( -3,0 ) ( 10,10)@xmath1 ( 55,0 ) ( 67,5)@xmath37 + c = 0.9 mm ( 140,20 ) ( 0,-2 ) ( 7,1)@xmath52 ( 55,-2 ) ( 57,7)@xmath53 ( 110,-2 ) ( 112,1)@xmath54 + the original cutting - plane method is due to gomory @xcite ( see also @xcite , ( * ? ? ?
* chapter14 ) ) .
gomory s method is applied to the problem in the case @xmath55 and can be formulated in terms of the functionals @xmath4 such that @xmath1 is an @xmath24-free set which is an infinite ` slab ' ( that is , the boundary of @xmath1 consists of two parallel hyperplanes ) .
balas @xcite introduced cutting - plane methods with respect to general @xmath24-free sets @xmath1 . for recent results related to cutting - plane methods based on general @xmath24-free sets we refer to @xcite
the @xmath5-closure operation @xmath6 can be used to estimate the ` cutting quality ' of methods which generate cuts from sets @xmath56 .
furthermore , finite resp .
infinite convergence properties of the sequences @xmath57 are related to finite resp .
infinite convergence of cutting - plane methods based on @xmath5 . with a view toward algorithmic applications , it is natural to ask for a ` convenient ' description of @xmath4 resp .
@xmath6 and for conditions on @xmath1 resp .
@xmath5 under which such a description can be given by finite data .
in particular , it is important to know whether for a given @xmath5 the functional @xmath6 maps polyhedra to polyhedra ( or rational polyhedra to rational polyhedra ) .
the above questions about the properties of @xmath4 and @xmath6 were addressed by andersen , louveaux and weismantel @xcite .
the convergence of sequences of the form @xmath58 was studied by owen and mehrotra @xcite and del pia and weismantel @xcite . in this manuscript
we generalize a part of results given in @xcite and @xcite .
furthermore , we also present new results which serve as a natural supplement .
in contrast to @xcite we do not restrict considerations to the classes @xmath5 consisting of @xmath24-free sets only . in most of the cases our assumptions on @xmath5 do not involve any mixed - integer space @xmath24 .
such a more general setting might be of interest for cutting - plane theory , since it seems possible that cutting - plane methods based on sets which are not necessarily @xmath24-free can also be introduced . in the purely integer case @xmath55 , sets @xmath1 with a fixed positive number of interior integer points could be a natural choice ( see also results from @xcite on the geometry of such sets ) .
for example , one can consider a set @xmath1 with precisely one interior integer point @xmath59 . with this choice , a cut generated by @xmath1 removes at most one point of @xmath35 , namely the point @xmath59 ( we recall that @xmath37 is the polyhedron defined by )
. a possible design template would be that a cutting - plane method based on sets @xmath1 as described above keeps the track of the ` best ' point removed , that is , the point @xmath59 with maximal @xmath60 among all points which were removed from @xmath35 during the execution of the method .
the aim of the method would be to change @xmath37 iteratively finally arriving at the situation @xmath61 or at the situation where @xmath62 yielding @xmath63 and belonging to @xmath24 can be found .
having riched such a situation , the method can easiliy determine the optimal solution ( which is either the best point that was removed or the point @xmath64 ) .
let us give an overview of the main results of the manuscript .
in theorem [ descr : r ] we study the properties of @xmath3 .
in particular , we describe the set of extreme points and the recession cone of @xmath3 . in theorem
[ polyhedrality : characterization ] we characterize those @xmath1 for which @xmath4 maps polyhedra to polyhedra . in theorem
[ polyhedrality : of : closures ] we consider @xmath5 consisting of rational polyhedra and present a condition on @xmath5 under which @xmath6 maps rational polyhedra to rational polyhedra .
even more generally , theorem [ polyhedrality : of : closures ] asserts that , under certain assumptions on @xmath5 , the set @xmath65 can be ` finitely generated ' , that is , @xmath66 for some finite subclass @xmath67 of @xmath5 .
the study of @xmath6 in the case of @xmath5 consisting of maximal @xmath24-free sets is of particular importance .
therefore , in theorem [ max - fac - width - of - max - lat - free ] we give a simple formulation of the condition on @xmath5 appearing in theorem [ polyhedrality : of : closures ] for this particular case .
theorems [ descr : r ] and [ polyhedrality : of : closures ] generalize the main results of @xcite , while theorems [ polyhedrality : characterization ] and [ max - fac - width - of - max - lat - free ] are ( to the best of our knowledge ) new . in theorem
[ convergence : thm ] we describe the limit of @xmath23 , as @xmath68 , under some weak assumptions on @xmath0 and @xmath5 .
theorem [ convergence : thm ] is related to theorem 2 from @xcite .
corollary [ convergence : bounded : case ] , which follows from theorem [ convergence : thm ] , is a convergence result extending theorem 3 from @xcite .
our proofs use standard tools of affine convex geometry ( facial structure of convex sets , recession cones and duality ) . in the proof of theorem [ polyhedrality : of : closures ] we use the gordan - dickson lemma .
the manuscript is organized as follows .
section [ prelim ] gives necessary preliminary information , section [ sect : results ] contains the formulations of the main results and section [ sect : proofs ] presents the proofs .
for information on convex geometry we refer to @xcite . the elements of @xmath2 are defined to be columns of @xmath18 real numbers .
the origin of @xmath2 is denoted by @xmath69 . by @xmath70
we denote the standard scalar product in @xmath2 .
we use the functionals @xmath71 ( affine hull ) , @xmath72 ( conical hull ) , @xmath16 ( convex hull ) , @xmath73 ( closure ) , @xmath17 ( interior ) , @xmath74 ( linear hull ) , @xmath75 ( @xmath18-dimensional volume ) . for @xmath76 , we define @xmath77:={\left\{(1-t ) a + t b \ , : \ , 0 \le t \le 1 \right\}}$ ] . if @xmath78 , then @xmath77 $ ] is called a _ segment _ in @xmath2 , and @xmath79 and @xmath32 are called the _ endpoints _ of @xmath77 $ ]
. we use @xmath80 to denote the _ relative interior _ of @xmath81 , i.e. , the interior of @xmath82 with respect to the euclidean topology of the affine space @xmath83 .
if @xmath84 is a convex set in @xmath2 , then the dimension @xmath85 of @xmath84 is defined to be the dimension of @xmath86 . for @xmath87 and @xmath88
we introduce @xmath89 for @xmath90 we also use the notations @xmath91 and @xmath92 . let @xmath0 be a nonempty closed convex set in @xmath2 .
with @xmath0 we associate the following functions and sets .
@xmath93 above @xmath94 .
the gauge function @xmath95 and the polar set @xmath96 are introduced under the assumption @xmath97 .
the values of the functions @xmath98 lie in @xmath99 .
it is known that the definitions and remain unchanged if one replaces the existential quantifier over @xmath14 by the universal quantifier .
one has @xmath100 where and are stated under the assumption @xmath97 .
we shall use the following simple proposition .
[ ray - prop ] let @xmath0 be a closed convex set in @xmath2 with @xmath101 .
let @xmath102 .
consider the ray @xmath103 .
then @xmath104 and , furthermore , the following statements hold .
a. one has @xmath105 if and only if the ray @xmath106 is contained in @xmath0 .
b. if @xmath107 , then @xmath108 is the unique point of the intersection of @xmath109 and @xmath106
. the following result can be found in ( * ? ? ?
* theorem 1.1.13 ) and ( * ? ? ? * theorem 6.9 ) .
[ char : relint ] let @xmath110 , @xmath111 and @xmath112 .
then @xmath113 if and only if there exist @xmath114 with @xmath115 and @xmath116 .
the following is a version of carathodory s theorem . .
[ strong : caratheodory ] let @xmath81 and let @xmath117 . then there exists an affinely independent set @xmath118 such that @xmath119 .
theorem [ strong : caratheodory ] is a direct consequence of the standard carathodory theorem ( see ( * ? ? ?
* theorem 1.1.4 ) ) and theorem [ char : relint ] .
a point @xmath14 of a convex set @xmath84 is said to _ extreme _ if there exists no segment @xmath120 satisfying @xmath121 .
the set of all extreme points of @xmath84 is denoted by @xmath122 .
we emphasize that the notion of extreme point is introduced with respect to sets which are not necessarily closed . for a subset @xmath82 of @xmath2 one
has @xmath123 inclusion can be derived from theorem [ strong : caratheodory ] .
the following characterization of extreme points is given in ( * ? ? ?
* lemma 1.4.6 ) . .
[ cap : lemma ] let @xmath0 be a closed convex set in @xmath2 and @xmath124
. then @xmath125 if and only if for every open neighborhood @xmath126 of @xmath127 there exists a hyperplane @xmath128 such that @xmath14 and @xmath129 lie in different open halfspaces defined by @xmath128 .
( see also fig .
[ fig : cap : lemma ] . )
c = 1.1 mm ( 44,20 ) ( 4,-3 ) ( 26,12)@xmath127 ( 0,9)@xmath0 ( 5,16)@xmath128 ( 33,16)@xmath126 + a convex subset @xmath130 of a closed convex set @xmath131 is said to be a _ face _ of @xmath0 if for every segment @xmath106 lying in @xmath0 and satisfying @xmath132 one necessarily has @xmath133 .
every face is necessarily a closed set .
furthermore , directly from the definition it can be seen that if @xmath84 is a convex subset of @xmath0 such , @xmath130 is a face of @xmath0 and @xmath134 , then @xmath135 .
it is known that every closed convex set @xmath0 in @xmath2 is the disjoint union of the relative interiors of all faces of @xmath0 ( see ( * ? ? ? *
theorem 2.1.2 ) ) .
furthermore , if @xmath136 is a face of @xmath0 and @xmath137 is a face of @xmath136 , then @xmath137 is a face of @xmath0 . given an integer @xmath138 by @xmath139 we denote the union of all faces of @xmath0 of dimension at most @xmath13 .
the set @xmath139 is said to be the _
@xmath13-skeleton _ of @xmath0 .
one has @xmath140 .
the one - dimensional faces of @xmath0 which are rays are called _ extreme rays _ of @xmath0 . by @xmath141
we denote the union of all extreme rays of @xmath0 .
a closed convex set @xmath0 is said to be _ line - free _ if @xmath0 does not contain lines .
every closed convex set is a direct sum of a linear space and a line - free closed convex set . by this , in most cases there is no loss of generality in considering line - free closed convex sets only
. the following result can be found in ( * ? ? ?
* theorem 1.4.3 and corollary 1.4.4 ) . .
[ decomp : conv : sets ] let @xmath0 be a line - free closed convex set in @xmath2
. then the following equalities hold .
@xmath142 given @xmath143 and @xmath39 by @xmath144 we denote the closed euclidean ball of radius @xmath145 centered at @xmath14 .
for closed convex sets @xmath146 in @xmath2 the _ hausdorff distance _ @xmath147 of @xmath148 and @xmath149 is defined by @xmath150 in particular , one has @xmath151 if one of the two sets @xmath152 is empty and the other is not .
the hausdorff distance is a metric on the class of closed convex sets in @xmath2 ( for further information , see @xcite ) .
some of our results deal with convergence of sequences of closed convex sets .
the convergence of such sequences can be introduced in several ways ( see ( * ? ? ?
* , 21.vii and
29.vi ) ) .
under restriction to bounded closed convex sets , the convergence with respect to the hausdorff distance is the standard choice . in the class of closed convex sets ( that is , without boundedness restriction ) also other forms of convergence are natural .
a sequence @xmath153 of convex sets in @xmath2 is said to be _ decreasing _ ( in the nonstrict sense ) if @xmath154 for every @xmath155 . given a decreasing sequence @xmath153 as above the set @xmath156
is the limit of @xmath153 in the sense of the definition given in @xcite .
however , in general , @xmath0 is not necessarily the limit of @xmath153 with respect to the hausdorff distance .
in fact , take @xmath157 and @xmath158 ( where @xmath159 is the standard basis of @xmath160 ) .
then @xmath161 , but @xmath162 for every @xmath155 . the following lemma ( see ( * ? ? ?
* lemma 1.8.1 ) ) shows that under boundedness assumption examples as above do not exist . .
[ decr : convergence ] let @xmath153 be a decreasing sequence of nonempty compact convex sets in @xmath2 .
then @xmath163 is a nonempty compact convex set and @xmath164 converges to @xmath0 with respect to the hausdorff distance , as @xmath68 .
a subset @xmath37 of @xmath2 is called a _ polyhedron _
if @xmath37 is intersection of finitely many closed halfspaces ( thus , the empty set is also a polyhedron ) . in analytic terms
, @xmath165 is a polyhedron if and only if there exist @xmath166 and @xmath167 with @xmath27 such that @xmath168 ( in the case @xmath169 one has @xmath170 . ) bounded polyhedra are said to be _ polytopes_. a polyhedron @xmath37 in @xmath2 is called _ rational _ if @xmath37 can be given by with @xmath171 and @xmath172 .
a polyhedron @xmath37 is said to be _ integral _ if @xmath173 .
every integral polyhedron is necessarily also rational .
if @xmath37 is a rational polyhedron , then @xmath174 and @xmath175 are integral polyhedra .
if @xmath37 is a polyhedron given by such that @xmath176 , @xmath177 and @xmath175 is a linear space , then @xmath178 and one has @xmath179 where @xmath180 is an arbitrary vector in @xmath2 . if @xmath37 is a polyhedron , then @xmath181 is precisely the set of all vertices of @xmath37 and , for @xmath182 , @xmath183 is the union of all @xmath13-dimensional faces of @xmath37 .
one - dimensional faces of polyhedra are called _
edges_. every edge is either a segment or a ray . for a rational @xmath18-dimensional polyhedron @xmath37
we denote by @xmath184 the set of all vectors @xmath185 such that @xmath180 is an outer normal to a facet of @xmath37 and the components of @xmath180 are relatively prime integers .
one has @xmath186 the _ max - facet - width _ of a @xmath18-dimensional rational polyhedron @xmath37 is defined by @xmath187 where @xmath188 is the width function , which was defined by .
it is not hard to see that @xmath189 if and only if @xmath175 is a linear space . by @xmath190
we denote the set of all affine transformations @xmath31 in @xmath2 satisfying @xmath191 .
two sets @xmath82 and @xmath192 in @xmath2 are said to be _ @xmath193-equivalent _ if @xmath194 for some @xmath195 .
if @xmath196 is a class of sets in @xmath2 , then by @xmath197 we denote the set of equivalence classes on @xmath196 with respect to the @xmath193-equivalence .
the notion of max - facet - width is invariant under @xmath193 equivalence , that is , @xmath198 for every rational polyhedron @xmath37 in @xmath2 and every @xmath199 .
the following theorem is well - known ( see ( * ? ? ?
* theorem 2 ) and @xcite ) .
[ finiteness : and : volume : boundedness ] let @xmath5 be a class of @xmath18-dimensional integral polytopes in @xmath2 . then the set @xmath200 is finite if and only if @xmath201 .
a subset @xmath0 of @xmath2 is said to be _ lattice - free _ if @xmath0 is @xmath18-dimensional closed convex and @xmath202 .
a lattice - free set @xmath0 is said to be _ maximal lattice - free _ if @xmath0 is not properly contained in another lattice - free set .
proposition [ max - m - free - prop ] below shows that maximal @xmath24-free sets ( which were defined in the introduction ) can be described in terms of maximal lattice - free sets . .
[ max - m - free - prop ] let @xmath8 .
let @xmath24 be a mixed - integer space given by .
then a set @xmath37 in @xmath2 is maximal @xmath24-free set if and only if @xmath203 , where @xmath204 is a maximal lattice - free set in @xmath205 .
the proof of proposition [ max - m - free - prop ] is straightforward and is therefore omitted .
the following proposition presents well - known properties of maximal lattice - free sets ( see ( * ? ? ?
* propositions 3.1 and 3.3 ) ) . .
[ max - lat - free - prop ] let @xmath0 be a lattice - free set in @xmath2 . then the following statements hold . a. [ max - lat - free - description ] the set @xmath0 is maximal lattice - free if and only if @xmath0 is a polyhedron and the relative interior of each facet of @xmath0 contains a point of @xmath193 .
b. [ unbounded - max - lat - free ] if @xmath0 is maximal lattice - free and unbounded , then @xmath0 is @xmath193-equivalent to @xmath206 , where @xmath207 and @xmath208 is a maximal lattice free set in @xmath209 . in view of propositions [ max - m - free - prop ] and [ max - lat - free - prop ] , every
maximal @xmath24-free sets is also maximal lattice - free .
we shall need the following useful fact ( for far - reaching generalizations formulated in the framework of well - quasi orderings see @xcite ) . .
[ gordan : dickson ] let @xmath210 .
then there exists a finite subset @xmath211 of @xmath82 such that every @xmath212 satisfies @xmath213 for some @xmath214 .
in view of theorem [ decomp : conv : sets ] we are motivated to describe the extreme points and the recession cone of @xmath3 .
this is done in theorem [ descr : r ] under rather weak assumptions on @xmath0 and @xmath1 .
parts i and ii of theorem [ descr : r ] generalize ( * ? ? ?
* lemma 4.2 ) and ( * ? ? ?
* lemmas 2.3 and 2.4 ) , respectively .
[ descr : r ] .
let @xmath0 be a line - free closed convex set in @xmath2 and let @xmath1 be a @xmath18-dimensional closed convex set in @xmath2 .
let @xmath215 and assume @xmath216 .
then the following statements hold .
a. [ ext : r ] for a point @xmath39 the following conditions are equivalent .
a. one has @xmath217 . b. either @xmath218 or there exists a one - dimensional face @xmath106 of @xmath0 such that @xmath219 consists of two connected components @xmath220 and @xmath221 which satisfy @xmath222 and @xmath223 .
b. [ rep : r ] if @xmath224 is a linear space , then @xmath225 is closed , the recession cones of @xmath225 and @xmath0 coincide and , furthermore , one has @xmath226 part [ ext : r ] of theorem [ descr : r ] can be illustrated by fig .
[ r_l : illustration ] from the introduction . in fig .
[ r_l : illustration ] the set @xmath227 is a @xmath228-gon . it can be seen that the seven vertices of @xmath229 are the only points that satisfy condition ( ii ) from part [ ext : r ] .
it turns out that , for a general @xmath1 , the functional @xmath4 does not always map polyhedra to polyhedra .
the reason of this is that @xmath3 is not always a closed set when @xmath0 is closed .
see fig .
[ fig - two - orhtants ] for an example .
our next result characterizes the sets @xmath1 for which @xmath4 maps polyhedra to polyhedra .
c ( 80,30 ) ( 6,-3 ) ( 1,20)@xmath0 ( 30,0)@xmath1 ( 40,20)@xmath3 + [ polyhedrality : characterization ] .
let @xmath1 be a @xmath18-dimensional closed convex set in @xmath2 .
then the following conditions are equivalent .
a. [ polyh->polyh ] @xmath4 maps every polyhedron in @xmath2 to a polyhedron .
b. [ closed->closed ] @xmath4 maps every closed convex set in @xmath2 to a closed convex set . c. [ rec : lin : or : halfspace ] either @xmath1 is a halfspace or @xmath224 is a linear space . in theorem [ polyhedrality : of :
closures ] below we present a condition on @xmath5 under which @xmath6 maps rational polyhedra to rational polyhedra .
theorem [ polyhedrality : of : closures ] is a generalization of ( * ? ? ?
* theorem 4.2 ) .
[ polyhedrality : of : closures ] .
let @xmath37 be a rational polyhedron in @xmath2 .
let @xmath5 be a class of @xmath18-dimensional rational polyhedra in @xmath2 such that the following conditions are fulfilled .
a. there exists @xmath230 such that @xmath231 for every @xmath56 .
b. there exists @xmath232 such that for every @xmath56 and every facet @xmath130 of @xmath1 one has @xmath233 . then the following statements hold .
a. there exists a finite subclass @xmath67 of @xmath5 such that every @xmath56 satisfies @xmath234 for some @xmath235 .
b. one has @xmath66 for some finite subclass @xmath67 of @xmath5 . c. the set @xmath65 is a rational polyhedron .
part i is the main assertion of theorem [ polyhedrality : of : closures ] .
part i resembles the structure of the gordan - dickson lemma .
in fact , we use the gordan - dickson lemma in its proof . as a consequence of theorem [ polyhedrality : of : closures ] we obtain the following result from ( * ? ? ?
* theorem 4.2 ) .
[ polyhedrality : of : closures : cor ] let @xmath37 be a rational polyhedron in @xmath2 .
let @xmath5 be a class of @xmath18-dimensional rational polyhedra in @xmath2 such that the following conditions are fulfilled .
a. there exists @xmath230 such that for every @xmath56 one has @xmath231 .
b. every @xmath56 is a maximal lattice - free set
. then parts i - iii of theorem [ polyhedrality : of : closures ] are fulfilled .
below we explain the relation of corollary [ polyhedrality : of : closures : cor ] to the chvtal and split closure . for @xmath236 let @xmath237 denote the greatest common divisor of the components of @xmath180 . a set @xmath1 in @xmath2
is called a _ split _ if @xmath238 where @xmath239 , @xmath185 and @xmath240 ( for an illustration see the set @xmath53 from fig . [ ex - m - free ] in the introduction ) .
a rational polyhedron @xmath1 in @xmath2 is a split if and only if @xmath241 and @xmath242 , where @xmath243 are distinct parallel hyperplanes in @xmath2 both containing points of @xmath193 .
let @xmath37 be a rational polyhedron in @xmath2 .
then the set @xmath244 is called the _ chvtal closure _ of @xmath37 ( see @xcite ) . clearly , @xmath245 for @xmath5 consisting of all splits @xmath1 having the form @xmath246 with @xmath185 and @xmath240 .
thus , corollary [ polyhedrality : of : closures : cor ] implies that for every rational polyhedron @xmath37 in @xmath2 the set @xmath247 is a rational polyhedron ( see ( * ? ? ?
* theorem 23.1 ) ) . consider a mixed - integer space @xmath24 ( given by ) .
we call a subset @xmath1 of @xmath2 an _ @xmath24-split _ if @xmath248 , where @xmath50 is a _ split _ in @xmath205
. if @xmath5 is the class of all @xmath24-splits , then we call the functional @xmath249 the _ split closure _
( with respect to @xmath24 ) . by corollary
[ polyhedrality : of : closures : cor ] , for every rational polyhedron @xmath37 in @xmath2 the split closure @xmath250 of @xmath37 with respect to @xmath24 is a rational polyhedron ( see ( * ? ? ? * theorem 3 ) ) .
the following theorem shows that , if the condition ( b ) of corollary [ polyhedrality : of : closures : cor ] is fulfilled , then condition ( a ) can be reformulated in simple terms .
[ max - fac - width - of - max - lat - free ] let @xmath5 be a class of rational polyhedra in @xmath2 such that each @xmath56 is a maximal lattice - free set .
then the following conditions are equivalent .
a. there exists @xmath230 such that for every @xmath56 one has @xmath231 .
b. the set @xmath200 is finite .
our next result deals with convergence of @xmath23 , as @xmath68 .
since we consider a decreasing sequence of sets , the convergence is expressed as an intersection . .
[ convergence : thm ] let @xmath5 be a nonempty class of @xmath18-dimensional closed convex sets in @xmath2 such that for every @xmath56 the set @xmath224 is a linear space .
let @xmath251 then the following statements hold .
a. for every line - free closed convex set @xmath0 in @xmath2 one has @xmath252 b. if @xmath253 is a mixed - integer space ( that is , @xmath254 with @xmath24 given by ) , then for every line - free rational polyhedron @xmath37 in @xmath2 one has @xmath255 under the given assumptions , theorem [ convergence : thm ] can not be improved to a result on finite convergence as an example from ( * ? ?
* example 2 ) shows ( but see ( * ? ? ?
* theorem 4 ) for conditions sufficient for a finite convergence ) .
[ convergence : bounded : case ] let @xmath0 , @xmath5 and @xmath253 be as in theorem [ convergence : thm ] and let @xmath0 be bounded
. then @xmath23 converges to @xmath256 in the hausdorff distance .
corollary [ convergence : bounded : case ] extends theorem 3 from @xcite which states that the sequence of split closures @xmath257 converges to @xmath46 in the hausdorff metric , as @xmath68 , in the case that @xmath37 is a rational polytope in @xmath2 and @xmath24 is a mixed - integer space ( given by ) .
del pia and weismantel showed that the latter result from from @xcite also holds if @xmath37 is an arbitrary rational polyhedron ( see ( * ? ? ? * theorem 2 ) ) .
we emphasize that , in the case when no boundedness assumptions are made , the convergence with respect to the hausdorff distance is a stronger assertion than the convergence in the sense of part ii of theorem [ convergence : thm ] .
thus , compared to theorem 2 from @xcite , theorem [ convergence : thm ] provides an infinite - convergence result with weaker assumptions and a weaker assertion .
let @xmath258 . first we show the inclusion @xmath259 since @xmath260 , we have @xmath261 .
let us show that @xmath262 .
assume the contrary , that is there exists @xmath263 . by the cap lemma ( lemma [ cap : lemma ] ) applied to @xmath264 and the open neighborhood @xmath265 of @xmath14 there exists a closed halfspace @xmath41 such that @xmath266 and @xmath267 .
it follows that @xmath268 and by this @xmath269 . on the other hand
, @xmath14 is a point of @xmath264 not lying in @xmath41 , a contradiction to @xmath270 .
hence is fulfilled . _
part [ ext : r ] .
_ let @xmath217 .
we show that ( ii ) is fulfilled . by
, we have @xmath271 .
hence , by the _ separation theorem _ ( see ( * ? ? ?
* theorem 1.3.4 ) ) , there exists a hyperplane @xmath128 with @xmath272 and @xmath273 .
one has @xmath274 .
in fact , if @xmath275 , then there exist a segment @xmath276 such that @xmath277 .
but then @xmath278 , a contradiction .
let @xmath106 be the face of @xmath0 with @xmath279 .
we have @xmath280 .
this can be shown arguing by contradiction . if @xmath281 , then @xmath282 and @xmath283 contradicting @xmath274 . in the case @xmath284 , one has @xmath285 and thus @xmath218 . consider the case @xmath286 .
if @xmath287 , then @xmath288 . hence there exists a sufficiently small segment @xmath289 with @xmath290 .
then @xmath278 , a contradiction .
thus , @xmath291 .
the set @xmath219 consists of two connected components @xmath220 and @xmath221 .
if both @xmath220 and @xmath221 contain points of @xmath292 , it follows @xmath278 , a contradiction to the choice of @xmath14 .
thus , we can assume that @xmath220 or @xmath221 is a subset of @xmath265 .
without loss of generality let @xmath222 .
then , by convexity of @xmath1 , one has @xmath223 .
this shows ( i ) @xmath293 ( ii ) .
let us show ( ii ) @xmath293 ( i ) .
we consider an arbitrary @xmath39 satisfying ( ii ) . in the case
@xmath218 , using the definition of the notion of extreme point , we see that the conditions @xmath294 and @xmath285 imply ( i ) .
otherwise @xmath14 lies in the relative interior of a one - dimensional face @xmath106 of @xmath0 . by assumption , @xmath219 consists of two connected components @xmath220 and @xmath221 that satisfy @xmath222 and @xmath223 .
let us verify @xmath217 .
we argue by contradiction .
assume that @xmath14 is not an extreme point of @xmath225 , that is , there exist @xmath295 such that @xmath296 and @xmath297)$ ] . by theorem
[ strong : caratheodory ] for each @xmath298 there exists an affinely independent set @xmath299 such that @xmath300 . by theorem [ char : relint ]
we deduce that @xmath301 for @xmath302 .
furthermore , @xmath14 is not an extreme point of @xmath303 since @xmath297)$ ] and @xmath304 . in view of theorem
[ strong : caratheodory ] , there exists an affinely independent set @xmath305 with @xmath119 .
we thus have @xmath306 , @xmath121 and @xmath106 is a face of @xmath0 .
hence , using properties of faces , we obtain @xmath307 .
the affinely independent set @xmath192 lies in a one - dimensional convex set @xmath106 and satisfies @xmath308 .
it follows that @xmath309 , where @xmath310 and @xmath311 . but @xmath222 and thus @xmath312 , a contradiction to @xmath313 .
this shows @xmath217 and completes the proof of ( ii ) @xmath293 ( i ) .
_ part [ rep : r ] .
_ we assume that @xmath224 is a linear space .
let us show the equality @xmath314 .
inclusion @xmath260 implies @xmath315 . in order to verify the reverse inclusion
we show @xmath316 we choose arbitrary @xmath127 and @xmath180 as in .
consider the convex set @xmath317 .
let us show that @xmath318 is bounded .
assume the contrary .
since @xmath319 , we see there exists @xmath320 such that @xmath321 . since @xmath318 is unbounded , we have @xmath322 for every @xmath323 .
thus @xmath324 . on the other hand , by the convexity of @xmath1 , @xmath325 for every @xmath326 .
thus , @xmath327 , @xmath324 , @xmath328 .
it follows that @xmath224 is not a linear space , a contradiction to the choice of @xmath1 .
the latter shows that @xmath318 is bounded .
the boundedness of @xmath318 implies @xmath329 for all sufficiently large @xmath330 .
thus , every @xmath331 with @xmath323 is a convex combination of @xmath332 and @xmath333 with an appropriate @xmath330 .
it follows @xmath334 .
hence is fulfilled . since @xmath335 , implies @xmath336 .
thus we have verified @xmath337 .
let us show that @xmath225 is closed , that is , @xmath338 .
the inclusion @xmath339 is trivial . in view of for showing @xmath340
it is sufficient to verify the inclusions @xmath341 and @xmath342 .
the inclusion @xmath341 follows from . for showing @xmath342
we consider an arbitrary extremal ray @xmath106 of @xmath264 .
let @xmath14 be the endpoint of @xmath106 .
since @xmath14 is an extreme point of @xmath106 and @xmath106 is a face of @xmath264 , it follows that @xmath343 . then implies @xmath217 .
we represent @xmath106 by @xmath344 , where @xmath345 . since @xmath315 , by one has @xmath346 for every @xmath323 .
that is , @xmath347 .
we have verified @xmath340 .
thus , @xmath338 .
equality follows from .
equality is a consequence of and part i. we choose @xmath350 such that both @xmath180 and @xmath351 do not belong to @xmath224 .
consider an arbitrary @xmath352 .
let @xmath106 be the line through @xmath14 parallel to @xmath180 . by the choice of @xmath180 ,
the intersection @xmath353 is bounded .
hence @xmath354 .
we have shown @xmath355 .
the reverse inclusion is trivial .
[ projection : lemma ] let @xmath356 be closed convex sets in @xmath2 having the form @xmath357 and @xmath248 where @xmath208 and @xmath50 are closed convex sets in @xmath205 and @xmath358 are integers with @xmath359
. then @xmath360 .
let us show @xmath293 .
assume that is fulfilled .
if @xmath1 is a halfspace , is trivial .
thus , we assume that @xmath224 is a linear space .
we consider an arbitrary closed convex set @xmath0 and show that @xmath3 is closed .
the case @xmath361 is trivial .
thus , let @xmath362 .
if @xmath363 , then by lemma [ no : reduction : case ] , we have @xmath349 and thus @xmath3 is closed .
consider the case @xmath364 .
if @xmath0 is line - free the assertion follows from theorem [ descr : r].[rep : r ] .
if @xmath0 is not line - free , then choosing an appropriate coordinate system in @xmath2 we can assume that @xmath0 and @xmath1 can be given as in lemma [ projection : lemma ] with @xmath208 being line - free .
thus , this case can be reduced to the case of a line - free @xmath0 .
let us show @xmath293 .
assume that is fulfilled .
let @xmath37 be an arbitrary polyhedron in @xmath2 .
it suffices to consider the case @xmath365 . if @xmath366 we argue in the same way as in the proof of the implication @xmath293 and obtain @xmath367 .
if @xmath37 is line - free we apply theorem [ descr : r ] . by theorem [ descr :
r].[rep : r ] , @xmath229 is a closed set and @xmath368 .
thus , @xmath369 is a polyhedral cone . by theorem [ descr : r].[ext : r ] , every extreme point of @xmath229 is either a vertex of @xmath37 or lies in the relative interior of an edge of @xmath37 .
furthermore , theorem [ descr : r].[ext : r ] yields that the relative interior of every edge of @xmath37 contains at most one extreme point of @xmath229 .
thus , the number of extreme points of @xmath229 is finite .
taking into account , it follows that @xmath229 is a minkowski sum of a polytope and a polyhedral cone .
hence @xmath229 is a polyhedron .
the case that @xmath370 and @xmath37 is not line - free can be reduced to the case of line - free sets @xmath37 with the help of lemma [ projection : lemma ] .
let us verify the implications ( i ) @xmath293 ( iii ) and ( ii ) @xmath293 ( iii ) .
assume that ( iii ) is not fulfilled .
let us show that then ( i ) and ( ii ) are also not fulfilled .
we represent @xmath224 as @xmath371 , where @xmath82 is a linear space , @xmath84 is a line - free closed convex cone and @xmath372 .
one has @xmath373 . since ( iii )
is not fulfilled , either @xmath374 and @xmath375 or @xmath376 . in the case
@xmath374 , we choose an arbitrary @xmath377 and an arbitrary @xmath378 . in the case
@xmath376 we consider a two - dimensional linear space @xmath192 such that @xmath379 is two dimensional and choose @xmath380 to be a basis of @xmath192 such that @xmath381 .
we fix @xmath382 .
since @xmath383 and @xmath382 , there exists @xmath323 such that @xmath384 .
we define @xmath385 . for an illustration to the following arguments see fig .
[ fig - constructing - non - closed ] .
let @xmath215 . by theorem [ descr : r].[ext : r ] , we have @xmath386 .
it follows that @xmath387 since otherwise one had @xmath388 with both @xmath127 and @xmath389 belonging to @xmath225 , which would yield a contradiction to @xmath390 . on the other hand , for every @xmath391 on
has @xmath392 .
this is shown as follows . since @xmath393 , the set @xmath394 is bounded . furthermore , from the definition of @xmath0 it follows @xmath395 for every @xmath330 .
therefore one can choose a sufficiently large @xmath396 such that the point @xmath397 belongs to @xmath292 .
the latter implies @xmath398 since @xmath399 that is , @xmath400 is a convex combination of the points @xmath401 and @xmath402 both belonging to @xmath225 .
since @xmath403 is arbitrary , we conclude that @xmath127 belongs to the closure of @xmath225 but not to @xmath225 .
thus , assuming that ( iii ) is not fulfilled we have shown that ( ii ) is not fulfilled .
since @xmath0 is a polyhedron and @xmath225 is not a polyhedron , also ( i ) is not fulfilled .
[ integrality : lemma ] let @xmath1 be a @xmath18-dimensional rational polyhedron in @xmath2 with @xmath407 and let @xmath408 .
assume that there exist constants @xmath409 such that the following conditions are fulfilled .
the assumption @xmath415 implies that @xmath224 is a linear space .
thus , by and we have @xmath416 consider an arbitrary @xmath417 .
one has @xmath418 above follows from ( b ) and ( c ) , and follows from ( a ) .
thus , @xmath419 and @xmath420 .
we conclude that @xmath421 is divisible by @xmath422 . by this , taking into account , we derive part i. part ii follows from part i in view of .
let @xmath37 be a nonempty line - free rational polyhedron in @xmath2 and @xmath1 be a @xmath18-dimensional closed convex set in @xmath2 such that @xmath224 is a linear space .
below we introduce a matrix @xmath423 , which we call the _ remainder matrix _ of @xmath37 with respect to @xmath1 .
let @xmath424 be the set of all vertices and @xmath425 the set of all edges of @xmath37 .
for @xmath426 and @xmath427 with @xmath428 let @xmath429 be a vector with @xmath430 and such that the greatest common divisor of the components of @xmath431 is equal to one .
then @xmath432 where the entries @xmath433 are defined as follows . in the case
@xmath434 we let @xmath435 . in the case
@xmath428 we define @xmath433 by the following formula .
@xmath436 the definition of the remainder matrix is motivated by .
in fact , if we keep the polyhedron @xmath37 fixed and vary @xmath1 , then ( in view of ) the ` size ' of the set @xmath229 can be expressed in terms of the ` size ' of @xmath437 .
the matrix @xmath423 is a quantitative representation for the size of @xmath437 .
more precisely , if we consider @xmath426 and @xmath427 with @xmath438 , then the entry @xmath433 has the following interpretation .
the largest possible value for @xmath433 is @xmath439 , which indicates that @xmath440 remains in @xmath437 .
the smallest possible value @xmath441 indicates that @xmath440 and the whole edge @xmath442 incident to @xmath440 is ` cut off ' by @xmath1 completely .
in other cases one has @xmath443 , which indicates that @xmath440 does not remain in @xmath437 but a part of @xmath442 remains in @xmath437 , while the value @xmath433 represents how large this part of @xmath442 is .
see also fig .
[ fig - illustration - rem - mx ] for an illustration .
( 115,32 ) ( 0,8 ) ( 0,14)@xmath440 ( 10,11)@xmath442 ( 25,18)@xmath1 ( 2,0)@xmath439 ( 45,14)@xmath440 ( 55,11)@xmath442 ( 65,18)@xmath1 ( 43,0)@xmath441 ( 90,14)@xmath440 ( 95,11)@xmath442 ( 97,18)@xmath1 ( 83,0)@xmath444 ( 65,25)@xmath431 [ how : much : remains ] let @xmath37 be a nonempty rational line - free polyhedron in @xmath2 and let @xmath1 and @xmath50 be @xmath18-dimensional closed convex sets in @xmath2 such that @xmath224 and @xmath445 are linear spaces and the relation @xmath446 is fulfilled
. then @xmath447 .
in view of one has @xmath448 where @xmath425 denotes the set of all edges of @xmath37 .
the same equality also holds with @xmath50 in place of @xmath1 .
thus , it suffices to show that the relation @xmath446 implies @xmath449 for every @xmath427 .
consider an arbitrary edge @xmath427 .
if @xmath442 is a segment with endpoints @xmath440 and @xmath450 , then the following implications hold .
@xmath451 , \nonumber \\ & & & & & r_l(v , e)= \|u(v , e)\|_{l - v}>0 , \nonumber \\ & & & & & r_l(w , e ) = + \infty .
\label{part : cut : off}\end{aligned}\ ] ] the implications and have straightforward proofs .
the implication follows from proposition [ ray - prop ] .
if @xmath442 is a ray and @xmath440 is the endpoint of @xmath442 , then the following implications hold .
@xmath452 in the equality @xmath453 follows by applying theorem [ descr : r ] to the one - dimensional polyhedron @xmath442 .
in fact , by theorem [ descr : r ] , the recession cones of @xmath454 and @xmath442 coincide and @xmath440 is a vertex of @xmath454 . hence @xmath453 . the implication is trivial .
the implication follows directly from proposition [ ray - prop ] .
let @xmath442 be an arbitrary edge of @xmath37 .
we shall use the implication given above to show @xmath455 .
consider the case that @xmath442 is a segment and let @xmath440 and @xmath450 be the endpoints of @xmath442 . if both @xmath433 and @xmath456 are zero or both @xmath433 and @xmath456 are @xmath457 one has @xmath458 or @xmath459 and the inclusion @xmath460 is fulfilled .
otherwise , one of the two values @xmath461 is finite and strictly positive and the other one is @xmath457 and the same is also valid for @xmath50 in place of @xmath1 . without loss of generality
let @xmath462 .
then , by @xmath463 one has @xmath464 and hence @xmath433 and @xmath465 are positive and finite .
it follows @xmath466 .
representing @xmath454 and @xmath467 as in the implication we deduce @xmath455 . without loss of generality
we assume @xmath468 . if for some @xmath56 one has @xmath469 , then by lemma [ no : reduction : case ] one has @xmath367 .
thus , without loss of generality we can assume that @xmath470 for every @xmath56 .
it suffices to consider the case that @xmath37 is line - free .
in fact , if @xmath37 is not line - free , then applying an appropriate transformation from @xmath190 to @xmath37 and to all elements of @xmath5 we can assume that @xmath203 , where @xmath204 is a line - free rational polyhedron .
in view of lemma [ projection : lemma ] we can apply a canonical projection to @xmath37 and to all elements of @xmath5 passing to the case of a line - free polyhedron @xmath37 .
thus , assume that @xmath37 is line - free . in view of lemma [ integrality : lemma ] , for every @xmath56
each component of @xmath471 is either a nonnegative integer or @xmath457 .
let us apply the gordan - dickson lemma to the set of matrices @xmath471 , with @xmath56 , which are partially ordered by @xmath9 .
we conclude that one can choose a finite subclass @xmath67 of @xmath5 such that for every @xmath56 there exists @xmath235 with @xmath472 .
above we applied the gordan - dickson lemma in the case where some of the entries can be @xmath457 .
note that the standard form of the gordan - dickson lemma ( lemma [ gordan : dickson ] ) , in which the entries are assumed to belong to @xmath473 , can be easily extended to this slightly more general case . by lemma [ how : much : remains ] , for @xmath235 and @xmath56 the relation @xmath472 implies @xmath234 .
this shows part i. part ii is a straightforward consequence of part i. assumption ( a ) implies that for every @xmath56 the set @xmath224 is a linear space .
hence by theorem [ polyhedrality : characterization ] , for every @xmath56 , the set @xmath474 is a polyhedron . using theorem [ descr : r ]
we see that the polyhedra @xmath229 with @xmath56 are rational .
in fact , taking into account part i of theorem [ descr : r ] and the rationality of @xmath1 and @xmath37 , we have @xmath475 .
by part ii of theorem [ descr : r ] , we have @xmath476 . since @xmath175 is a rational polyhedron
, we deduce that @xmath229 is a minkowski sum of two rational polyhera .
it follows that @xmath229 is a rational polyhedron . for the finite set @xmath67 defined above we have @xmath477 .
thus , @xmath65 is intersection of finitely many rational polyhedra .
it follows that @xmath65 is a rational polyhedron , that is , part iii is fulfilled .
let us first give an example showing that assumption ( a ) alone is not enough .
let @xmath0 be the intersection of @xmath478 ^ 2 $ ] with a circular disk of radius @xmath479 centered at the origin ( thus , @xmath0 is a circular sector ) . for rational values @xmath480 and @xmath481 such that the line through @xmath482 and @xmath483 does not meet @xmath0 we define @xmath484 \times [ s,1]^d$ ] .
let @xmath5 be the set of all @xmath485 as described above ( see also fig .
[ fig - ex - b - necessary ] ) .
then @xmath5 satisfies the assumption ( a ) of theorem [ polyhedrality : of : closures ] since @xmath486 for every @xmath56 . by construction @xmath487 ^ 2)=k$ ] .
thus , @xmath488 ^ 2 $ ] is a polyhedron and @xmath487 ^ 2)$ ] is not a polyhedron .
let us give an example showing that assumption ( b ) alone is not enough .
again , let @xmath157 and let us use @xmath0 from the previous example . for every rational point
@xmath495^d \setminus k$ ] there exist points @xmath496 and @xmath497 such that @xmath498 , @xmath0 and @xmath127 lie in different open halfspaces defined by the line @xmath499 and , furthermore , both @xmath492 and @xmath493 do not belong to @xmath488 ^ 2 $ ] . for @xmath492 and @xmath493 as above
we define the triangle @xmath500 with vertices @xmath492 , @xmath493 and @xmath494 .
let @xmath5 be the set of all @xmath500 as described above ( see also fig .
[ fig - ex - a - necessary ] ) .
then @xmath5 satisfies the assumption ( b ) of theorem [ polyhedrality : of : closures ] ( with @xmath501 ) and @xmath487 ^ 2 ) = k$ ] .
thus , also for this example @xmath488 ^ 2 $ ] is a polyhedron and @xmath487 ^ 2)$ ] is not a polyhedron .
@xmath502 in view of proposition [ max - lat - free - prop].[max - lat - free - description ] , condition ( b ) of corollary [ polyhedrality : of : closures : cor ] implies condition ( b ) of theorem [ polyhedrality : of : closures ] with @xmath501 .
thus , the assumptions of theorem [ polyhedrality : of : closures ] are fulfilled and the assertion follows .
consider @xmath504 with @xmath505 let @xmath506 .
then there exist linearly independent @xmath507 such that @xmath508 .
for the parallelotope @xmath509 one has @xmath510 and @xmath511 .
since @xmath512 are linearly independent there exists @xmath513 such that @xmath514 for every @xmath12 . then @xmath515 let @xmath516 be the integer matrix with rows @xmath517 ( in this order ) . then @xmath518 can be given by @xmath519^d \right\ } } = b^{-1}([0,k]^d).\ ] ] thus @xmath520 the implication ( ii ) @xmath293 ( i ) follows from the invariance of the max - facet - width with respect to transformations from @xmath190 .
let us show ( i ) @xmath293 ( ii ) .
we assume ( i ) .
we represent @xmath5 as the union @xmath521 .
thus , it suffices to show the finiteness of @xmath522 for each @xmath523 . in view of proposition
[ max - lat - free - prop].[unbounded - max - lat - free ] it suffices to consider the case @xmath524 only .
that is , we restrict considerations to the case that for every @xmath56 the set @xmath1 is a @xmath18-dimensional polytope .
let @xmath56 . by proposition [ max - lat - free - prop].[max - lat - free - description ] the relative interior of each facet of the @xmath18-dimensional polytope @xmath1 contains an integral point ;
thus , the convex hull of such integral points is @xmath18-dimensional .
it follows that the set @xmath525 consists of @xmath18-dimensional integral polytopes . by ( i ) and
proposition [ vol : versus : maxfw ] the volume of the polytopes @xmath526 is bounded from above by @xmath527 .
hence also the volume of the polytopes @xmath528 is bounded by @xmath527 .
hence , by theorem [ finiteness : and : volume : boundedness ] , @xmath529 is finite .
we fix @xmath110 and polytopes @xmath530 such that @xmath531 for each @xmath532 we choose @xmath533 .
furthermore we fix @xmath534 such that @xmath535 for every @xmath532 .
consider an arbitrary @xmath56 . by construction
, there exists @xmath199 and @xmath532 such that for @xmath536 one has @xmath537 .
the inclusion @xmath538 implies @xmath539 where ( in view of and ) the polytope @xmath540 can be given by @xmath541 with @xmath542 for every @xmath543 one has @xmath544 hence @xmath545 is a natural number not larger than @xmath546 .
it follows that @xmath547 is divisible by @xmath548 and we deduce @xmath549 the latter implies @xmath550 . from @xmath551
and we derive @xmath552 .
thus , we have shown that for every @xmath56 there exists @xmath532 , a set @xmath553 and a transformation @xmath199 such that @xmath554 . since , for every @xmath532 , the set @xmath555 containing @xmath82 is finite , there exist only finitely many choices of sets @xmath82 as above .
this yields ( ii ) and finishes the proof of ( i ) @xmath293 ( ii ) .
[ boundedness : lemma ] let @xmath153 be a decreasing sequence of closed convex sets in @xmath2 ( that is , @xmath154 for every @xmath155 ) .
let @xmath163 be bounded and nonempty .
then there exists @xmath556 such that , for every @xmath155 with @xmath557 , the set @xmath164 is bounded .
we assume the contrary , that is , @xmath153 contains infinitely many unbounded sets .
since @xmath153 is decreasing it follows that , for every @xmath155 , the set @xmath164 is unbounded . for each @xmath155
we choose a unit vector @xmath558 in @xmath559 there exists an infinite subsequence @xmath560 of the sequence @xmath561 such that @xmath562 converges to some unit vector @xmath94 , as @xmath68 .
we fix an arbitrary @xmath124 .
consider arbitrary @xmath563 and @xmath323 . by construction
, one has @xmath564 for every @xmath556 with @xmath565 . taking limit , as @xmath566
, we obtain @xmath567 .
since @xmath13 is arbitrary we deduce @xmath568 .
since @xmath323 is arbitrary , it follows that @xmath569 .
thus @xmath0 is unbounded , a contradiction to the assumptions .
it is known that @xmath572 is a rational polyhedron ( see , for example , @xcite ) .
the inclusion @xmath573 is trivial .
let us show the reverse inclusion .
the cone @xmath175 is an integral polyhedron .
we represent @xmath175 by @xmath574 , where @xmath230 and @xmath575 . since @xmath576 , for all @xmath577 , @xmath578 and @xmath579 one has @xmath580 .
since @xmath581 , the latter shows @xmath582 for every @xmath579 and yields @xmath583 . this shows the assertion .
let @xmath0 be an arbitrary line - free closed convex set in @xmath2 and let @xmath584 .
in view of theorem [ descr : r].[rep : r ] , @xmath23 is closed for every @xmath138 .
hence , also @xmath225 is closed .
we show @xmath585 assume is not fulfilled .
then one can find @xmath586 with @xmath587 .
since @xmath587 , we can choose @xmath56 with @xmath382 .
we also choose a bounded open set @xmath126 satisfying @xmath588 . by the cap lemma ( lemma [ cap : lemma ] )
there exists a hyperplane @xmath128 such that @xmath127 and @xmath589 lie in different open halfspaces defined by @xmath128 .
let @xmath590 and @xmath41 be the closed halfspaces defined by @xmath128 and satisfying @xmath591 and @xmath592 . by the choice of @xmath128 , we have @xmath593 and thus , @xmath594 is bounded . since @xmath595
, applying lemma [ boundedness : lemma ] we obtain the existence of @xmath155 such that for every @xmath556 with @xmath596 the set @xmath597 is bounded . by lemma [ decr : convergence ] , the sequence @xmath598 converges to @xmath594 in the hausdorff distance .
we recall that for @xmath599 and @xmath39 by @xmath144 we denote the closed ball of radius @xmath145 centered at @xmath14 .
we can choose @xmath599 such that @xmath600 in view of the convergence of @xmath597 to @xmath594 and by the definition of the hausdorff distance , there exists @xmath155 such that @xmath601 with @xmath13 as above , using and we derive @xmath602 thus , @xmath603 .
it follows @xmath604 we derived the inclusion @xmath605 which contradicts @xmath606 .
this shows .
let us verify part i. we notice that if @xmath607 , then part i is fulfilled , since in this case one has @xmath608 .
this can be shown by contradiction .
assume that @xmath607 and @xmath216
. then @xmath609 .
by we get @xmath610 .
thus , @xmath611 , a contradiction .
therefore for the rest of the proof we assume @xmath611 .
for every integer @xmath138 we have : @xmath612 condition can be verified for every integer @xmath138 by induction on @xmath13 . for @xmath524
the condition trivially holds .
if the condition holds for some integer @xmath138 , then ( since @xmath224 is a linear space ) we can apply theorem [ descr : r].[rep : r ] .
it follows that @xmath613 is closed and one has @xmath614 for every @xmath56 .
consequently @xmath615 is closed and one has @xmath616 .
this yields for every integer @xmath138 . from and the definition of @xmath225 we deduce that @xmath225 is closed and one has @xmath617 . using the above observations and the inclusion
@xmath618 we derive @xmath619 .
the reverse inclusion is obtained by representing @xmath225 with the help of and applying together with @xmath620 : @xmath621 this shows part i. let us show part ii .
assume that @xmath253 is a mixed - integer space and let @xmath37 be a rational line - free polyhedron in @xmath2 .
the case @xmath622 follows directly by part i. therefore we assume @xmath623 . by lemma [ rec :
mi : hull ] and part i we obtain @xmath624 .
this finishes the proof of part ii . by theorem [ convergence :
thm].ii one has @xmath625 .
since the sequence @xmath58 consists of bounded closed convex sets and is decreasing , applying lemma [ decr : convergence ] , we see that @xmath23 converges to @xmath626 , as @xmath68 , with respect to the hausdorff distance .
k. andersen , q. louveaux , r. weismantel , and l. a. wolsey , _ inequalities from two rows of a simplex tableau _ , integer programming and combinatorial optimization , lecture notes in comput .
4513 , springer , berlin , 2007 , pp .
mr 2011b:90082 to3em , _ disjunctive programming : cutting planes from logical conditions _ , nonlinear programming , 2 ( proc . sympos
. special interest group on math .
programming , univ .
wisconsin , madison , wis . , 1974 ) , academic press , new york , 1974 , pp .
mr 51 # 12328 s. s. dey and l. a. wolsey , _ lifting integer variables in minimal inequalities corresponding to lattice - free triangles _ , integer programming and combinatorial optimization , lecture notes in comput .
5035 , springer , berlin , 2008 , pp .
mr 2011a:52014 r. g. jeroslow , _ cutting - plane theory : disjunctive methods _ , studies in integer programming ( proc .
workshop , bonn , 1975 ) , north - holland , amsterdam , 1977 , pp .
293330 . ann . of discrete math .
1 . mr 56 # 10953 h. marchand , a. martin , r. weismantel , and l. wolsey , _ cutting planes in integer and mixed integer programming _ , discrete appl . math .
* 123 * ( 2002 ) , no . 1 - 3 , 397446 , workshop on discrete optimization , do99 ( piscataway , nj ) .
mr 2003g:90045 g. nemhauser and l. wolsey , _ integer and combinatorial optimization _ , wiley - interscience series in discrete mathematics and optimization , john wiley & sons inc .
, new york , 1999 , reprint of the 1988 original , a wiley - interscience publication .
mr 2000c:90001 m. pouzet , _ applications of well quasi - ordering and better quasi - ordering _ , graphs and order ( banff , alta . , 1984 ) ,
147 , reidel , dordrecht , 1985 , pp . | for convex sets @xmath0 and @xmath1 in @xmath2 we define @xmath3 to be the convex hull of all points belonging to @xmath0 but not to the interior of @xmath1 .
cutting - plane methods from integer and mixed - integer optimization can be expressed in geometric terms using functionals @xmath4 with appropriately chosen sets @xmath1 .
we describe the geometric properties of @xmath3 and characterize those @xmath1 for which @xmath4 maps polyhedra to polyhedra . for certain natural classes @xmath5 of convex sets in @xmath2 we consider the functional @xmath6 given by @xmath7 .
the functional @xmath6 can be used to define various types of closure operations considered in the theory of cutting planes ( such as the chvtal closure , the split closure as well as generalized split closures recently introduced by andersen , louveaux and weismantel ) .
we study conditions on @xmath5 under which @xmath6 maps rational polyhedra to rational polyhedra .
we also describe the limit of the sequence of sets obtained by iterative application of @xmath6 to @xmath0 .
a part of the presented material gives generalized formulations and unified proofs of several recent results obtained by various authors .
primary 90c11 ; secondary 52a20 , 52b20 , 90c10 .
chvtal closure ; cutting plane ; disjunctive programming ; gomory cut ; max - facet - width ; mixed - integer optimization ; recession cone ; split closure ; width . |
the magneto - optical properties of a two - dimensional electron gas ( 2deg ) systems have been intensively studied experimentally and theoretically .
it is known that in a dilute system confined in a so - called symmetric quantum well ( qw ) , the photoluminescence ( pl ) spectrum is determined by a charged - exciton complex @xmath4 ( bound state of two electrons @xmath6 and a valence hole @xmath7 ) and its interaction with the remaining electrons .
the existence of @xmath4 in bulk semiconductors was predicted by lampert @xcite , but only in 2d quantum wells ( qw s ) does its binding energy turn out sufficiently large@xcite for the experimental detection .
the observation of the @xmath4 stimulated theoretical work , and it is now established that the only bound @xmath4 state at zero magnetic field is the singlet state ( @xmath8 ) with the total electron spin @xmath9 .
accordingly , the pl spectrum usually shows two peaks , due to the @xmath3 ( neutral exciton ) and/or @xmath8 recombination , split by the @xmath8 binding energy @xmath10 .
the situation is more complicated in a magnetic field @xmath11 . at very high @xmath11 ,
the optically active @xmath3 s decouple from the electrons due to the `` hidden symmetry''@xcite and the @xmath8 unbinds .
interestingly , a different @xmath4 state binds in high fields@xcite .
it is a triplet ( @xmath12 ) with @xmath13 and finite total angular momentum .
it has infinite radiative lifetime @xmath14@xcite because of ( independently ) the `` hidden '' and 2d translational symmetry , and thus will be further called @xmath15 ( @xmath0 for `` dark '' ) .
although both symmetries are broken in experimental systems by mixing of landau levels ( ll s ) , valence band mixing effects , asymmetry of the qw , and disorder , the @xmath15 recombination is expected to be weak and disappear at very large @xmath11 .
the fact that @xmath15 unbinds at @xmath16 while @xmath8 unbinds at high @xmath11 implies a singlet
triplet crossing , which in a gaas qw of width @xmath17 nm was predicted at @xmath18 t@xcite .
surprisingly , the following pl experiment@xcite showed no such transition up to @xmath19 t. this puzzle was resolved by the identification of another bound state , a radiative excited spin - triplet @xmath20 ( @xmath21 for `` bright '' ) @xcite . the emission spectrum from both triplets was eventually measured @xcite , and agrees well with the theoretical prediction@xcite .
the fact that the pl spectrum of a 2deg in a symmetric qw only measures the @xmath3 and/or @xmath4 emission means that it is not a useful probe of electron correlations .
the effect of surrounding 2deg on the @xmath4 emission is specially weak in dilute systems due to the laughlin correlations effectively isolating an @xmath4 from the electrons@xcite .
indeed , the measured pl spectra are remarkably insensitive to the electron density@xcite , and only at filling factors @xmath22 approaching @xmath23 does the relative position of pl peaks change with density . the pl spectra containing more information about the original ( laughlin ) electron correlations of the 2deg are obtained in asymmetrically doped wide qw s . in such structures , the spatial separation @xmath0 of @xmath6 and @xmath7 layers weakens the @xmath6@xmath7 interaction@xcite , and the pl spectra show discontinuities@xcite at the filling factor @xmath24 at which laughlin incompressible liquid state@xcite is formed and the fractional quantum hall ( fqh ) effect@xcite is observed in transport experiments .
the reason for these discontinuities is that at @xmath0 exceeding the magnetic length @xmath1 , the effective coulomb potential of the hole seen by the electrons is too weak and its resolution is too low , and a hole can no longer bind `` whole '' electrons to form @xmath4 s . instead , the hole interacts with charge excitations that are already present in the 2deg and which near @xmath24 happen to be a small number of laughlin quasiparticles ( qp s ) . at @xmath25 ,
the hole repels positively charged quasiholes ( qh s ) and thus remains in the locally undisturbed incompressible electron state ( such state can be called a `` decoupled hole '' and denoted by @xmath7)@xcite .
the situation is quite different at @xmath26 , when the hole attracts negatively charged quasielectrons ( qe s ) and binds one or two of them to form a fractionally charged exciton ( fcx ) , @xmath7qe or @xmath7qe@xmath27@xcite . in this paper , we discuss results of numerical calculations of various radiative @xmath6@xmath7 complexes that contribute to the pl spectrum of a 2deg depending on @xmath0 and @xmath22 .
the energy and pl spectra are obtained from exact diagonalization in haldane spherical geometry@xcite .
the presented unified description of the pl from the 2deg at high @xmath11 given in terms of emission from well defined bound states ( @xmath3 , @xmath4 , and @xmath7qe@xmath28 ) reduces the problem of a very complicated interaction between a hole and the 2deg to a simpler one , of determining the single - particle properties of the bound `` quasiparticles '' .
in particular , the optical selection rules for these states are formulated , following from the translational invariance .
preserving the 2d translational symmetry of an infinite 2deg in a finite - size calculation turns out to be essential for the identification of the optical selection rules of bound @xmath6@xmath7 states and the degeneracies in their energy spectrum . for any system of electrons and/or holes ,
this symmetry causes conservation of two orbital quantum numbers . for systems with total electric charge , @xmath29 ( such as @xmath4 ) , they are the total angular momentum @xmath30 and an additional angular momentum quantum number @xmath31 associated with a partial decoupling of the cm motion in a homogeneous magnetic field@xcite .
the energy levels of a charged system fall into degenerate ll s labeled by @xmath32 , and the states within each ll have @xmath33 , 1 , 2 , .
since both @xmath30 and @xmath31 commute with the pl operator @xmath34 , the following selection rule governs recombination of isolated charged @xmath6@xmath7 complexes : @xmath35 . in order to confine electrons to a finite area ( in order to achieve finite degeneracy of electron and hole ll s ) without breaking the 2d symmetry , haldane @xcite proposed to put electrons and holes on a sphere of radius @xmath36 . the magnetic field
@xmath11 perpendicular to the surface is due to a monopole placed in the center . the monopole strength @xmath37 is defined in the units of elementary flux @xmath38 , so that @xmath39 and the magnetic length is @xmath40 .
the single - particle states are the eigenstates of angular momentum @xmath41 and its projection @xmath42 and are called monopole harmonics @xcite .
the energies @xmath43 fall into @xmath44-fold degenerate angular momentum shells separated by the cyclotron energy @xmath45 .
the @xmath46-th ( @xmath47 ) shell ( ll ) has @xmath48 and thus @xmath37 is a measure of the system size through the ll degeneracy . due to the spin degeneracy ,
each shell is further split by the zeeman gap , @xmath49 . as a result of 2d rotational invariance ,
a many - body @xmath6@xmath7 system on a sphere has two good quantum numbers , length @xmath50 and projection @xmath51 of the total angular momentum @xmath52 .
the mapping between quantum numbers @xmath30 and @xmath31 on a plane and the 2d algebra of @xmath52 on a sphere allows conversion of the results between the two geometries@xcite .
in particular , ll s of a charged @xmath6@xmath7 complex are represented on a sphere by @xmath50-multiplets , states within each ll are labeled by @xmath51 , and the optical selection rule is @xmath53 .
the price paid for closing the many - body hilbert space without breaking the 2d symmetry is the surface curvature affecting interaction matrix elements .
however , if the correlations modeled have short range @xmath54 , the effects of curvature ( scaled by a small parameter @xmath55 ) can be eliminated by extrapolation to @xmath56 . using a composite index
@xmath57 $ ] ( @xmath58 is the spin projection ) , the @xmath6@xmath7 hamiltonian can be written as @xmath59 , where @xmath60 and @xmath61 create and annihilate particle @xmath62 ( @xmath6 or @xmath7 ) in state @xmath63 , and @xmath64 are the coulomb matrix elements .
the hamiltonian @xmath65 is diagonalized numerically in the basis of @xmath66@xmath67 slater determinants
. small density of holes is assumed which allows ignoring @xmath7@xmath7 interaction effects and inclusion of only one valence hole in the basis .
the @xmath66@xmath67 eigenstates are labeled by @xmath50 , @xmath51 , and @xmath68 .
the most important factors that determine dimension of the hamiltonian matrix that must be diagonalized are the number of electrons @xmath69 , the ll degeneracy controlled by @xmath37 , the number of excited ll s included , and the inclusion of spin - unpolarized states . using modified lanczos algorithms we diagonalized hamiltonians of dimensions up to about @xmath70 . in the @xmath4 problem , the number of particles is small and the inclusion of reversed spins and ll s is possible . the numerical results for @xmath4 s agree well with recent experiments @xcite .
however , when studying fcx s , as many electrons as possible need be included in the calculation , and the model must be simplified by assuming maximum spin polarization ( @xmath71 ) , neglecting excited ll s , and setting zero qw width , @xmath72 .
this allowed for calculation for @xmath73 at @xmath74 , but the results are largely qualitative .
the potential @xmath75 seen by the electrons due to the positive charge of the optically injected valence hole @xmath7 can be described by two effective parameters , strength @xmath76 and spatial resolution @xmath77 .
the response of the 2deg to the perturbation @xmath75 depends on the relation between @xmath76 and @xmath78 and the characteristic energies and lengths of the unperturbed system .
the pair of perturbation parameters @xmath76 and @xmath77 in principle can be varied independently , but in experimental samples they are predominantly controlled through the effective spatial separation @xmath0 of the hole from the 2deg layer .
the magnitude of @xmath0 depends on the electric field oriented across the 2deg plane and resulting from asymmetric doping . in symmetrically doped qw s ,
the hole moves in the same physical layer as the electrons ( @xmath79 ) .
however , in the asymmetrically ( one - sided ) doped qw s the displacement @xmath0 can be as large as the qw width @xmath80 .
clearly , at @xmath79 both @xmath76 and @xmath77 are the largest , and the strongest response of the 2deg to the hole can be expected .
the actual magnitudes of @xmath76 and @xmath77 at @xmath79 depend on the details of hole wave function , which to some extent depend on the magnetic field . in the extreme case of @xmath81 ( lowest ll approximation ) , the @xmath6 and @xmath7 wave functions are identical and the interaction hamiltonian is particle
hole symmetric@xcite .
this `` hidden '' symmetry causes vanishing of all electric moments in the optically active ground state of a bound @xmath6@xmath7 pair ( @xmath3 ) , which in turn causes the decoupling of @xmath3 s from the 2deg . as a result
, the @xmath3 is the most strongly bound optically active @xmath6@xmath7 state that can form in the 2deg . in the opposite case ,
the ( infinitely heavy ) hole acts as a point charge and its attraction to an electron is larger by a factor of @xmath82 . in an intermediate situation
, the enhanced @xmath6@xmath7 attraction always breaks the hidden symmetry and causes coupling of an @xmath3 to electrons . as a result , radiative @xmath4 states occur , and the @xmath4 rather than @xmath3 should be regarded the most stable bound state formed by a hole injected into the 2deg@xcite .
this can be rephrased as that at finite @xmath11 and @xmath79 the 2deg responds to @xmath75 by binding two electrons to the hole to form an @xmath4 and screen the hole s positive charge . in the other limit of large @xmath0 ,
both @xmath76 and @xmath77 are too small to cause a strong response of an incompressible electron liquid .
a hole is no longer able to pick out and bind a single electron from the 2deg , which can be understood by noticing that the elementary charge excitations of an unperturbed 2deg ( qe s and qh s ) have finite energy ( @xmath83 and @xmath84 ) , and that the electron wave function in the vicinity of the hole that corresponds to an @xmath4 can be expanded in the basis of these qe and qh excitations . on the other hand , too large @xmath0 implies too large @xmath78 and the size of an isolated @xmath4 ( @xmath85@xmath67 ground state ) that would exceed the characteristic @xmath6@xmath6 distance @xmath86 ( @xmath87 means density ) and make such an @xmath4 unstable when inserted into the 2deg ( independently of the preceding energetic argument ) .
the only allowed response of the 2deg to the hole in this ( `` weak '' ) coupling regime is to screen the hole s positive charge with the negatively charged qe s that are already present in the 2deg .
the fact that only the existing qe s can bind to the hole makes the response critically depend on the filling factor @xmath22 .
for example , near the laughlin filling of @xmath24 , the hole repels the qh s and moves in the locally incompressible @xmath24 liquid , causing no local response of the 2deg .
conversely , the binding of one or more qe s to the hole is expected at @xmath26 , and the resulting bound fcx states @xmath7qe@xmath28 are the most stable bound states at large @xmath0 . since the `` uncoupled '' state @xmath7 has different emission energy @xmath88 and oscillator strength @xmath89 than the @xmath7qe@xmath28 s , discontinuity in the pl spectrum
is expected at @xmath24 , in agreement with the experimental pl data for asymmetric structures @xcite .
[ [ isolated - charged - exciton - energy - spectrum . ] ] isolated charged exciton : energy spectrum .
an isolated neutral ( @xmath3 ) or charged ( @xmath4 ) exciton consists of only two or three particles . in the numerical calculation of the energy and pl spectra of this relatively simple quantum mechanical system we were able to include the effects of the finite magnetic field @xmath11 that causes ll mixing ( by including up to five electron and hole ll s , @xmath90 , with the lowest ll degeneracy of @xmath91 ) and of the finite qw width @xmath80 ( by using the values of the two - body coulomb matrix elements appropriate for the lowest subband of the qw rather than for an ideal 2d system ) . it should be stressed that due to different electron and hole effective masses and due to a different height of the electron and hole confinement potential at the well / barrier interface , the electron and hole wave functions in the @xmath92-direction are different even in the symmetric structures with @xmath79 . to a first approximation ,
this is included by introducing a pair of effective widths of the electron and hole layers , @xmath93 and @xmath94 , obtained by fitting the actual density profile @xmath95 of each carrier with @xmath96 . for a @xmath17 nm gaas / al@xmath97ga@xmath98as qw the effective widths are @xmath99 nm and @xmath100 nm .
the zeeman energy @xmath101 ( @xmath102 is the effective giromagnetic factor and @xmath103 is the bohr magneton ) enters the problem of an @xmath4 in two ways .
( i ) because of the complete spin polarization of the 2deg , binding of an @xmath4 state may or may not require a spin flip , depending on the total spin @xmath68 of two electrons in the bound @xmath4 state .
therefore , the binding energy @xmath104 of each spin triplet ( @xmath13 ) @xmath4 state is insensitive to the zeeman energy , but for spin singlet ( @xmath9 ) states @xmath104 is reduced by @xmath105 compared to the pure coulomb binding energy . for gaas , @xmath105 is roughly a linear function of energy through both cyclotron energy @xmath106 and confinement energy @xmath107 .
after snelling et al.@xcite , for @xmath108 nm at @xmath16 , we have @xmath109 with @xmath110 nm and @xmath111 .
after seck et al.@xcite we find @xmath112 t@xmath113 .
( ii ) the recombination of an @xmath3 or @xmath4 state can occur through relaxation of a conduction electron with either spin to the valence band .
since the angular momentum of the photon ( @xmath114 ) depends on the spin of the relaxing electron , and the emission energy @xmath88 includes the zeeman energy of the recombining @xmath6@xmath7 pair , @xmath115 , each pl peak associated with the emission from an @xmath3 or @xmath4 state is split into two circularly polarized peaks by @xmath115 .
the splitting is most evident for @xmath8 because of the large population of both electron spin levels even at a very low temperature . in fig .
[ fig1 ] we show the @xmath85@xmath67 energy spectra calculated for a @xmath116 nm gaas qw at @xmath117 , 30 , and 68 t. = 4.8 in in each frame , the energy @xmath118 is measured from the exciton energy @xmath119 , so that for the bound @xmath4 states it is opposite to the binding energy @xmath104 .
open and full symbols denote singlet and triplet electron spin configurations , respectively , and only the state with the lowest zeeman energy ( @xmath120 ) is marked for each triplet .
similarly , each state with @xmath121 represents a degenerate multiplet with @xmath122 .
the angular momentum @xmath50 calculated in the spherical geometry and given on the horizontal axes under each graph translates into the angular momentum on a plane @xmath123 in such way that the @xmath124 and @xmath125 multiplets correspond to the planar ll s with @xmath126 and @xmath127 , respectively
. [ [ isolated - charged - exciton - radiative - recombination . ] ] isolated charged exciton : radiative recombination . due to the conservation of @xmath50 in the pl process ,
only those @xmath85@xmath67 states from the @xmath124 channel are radiative .
this is because@xcite an annihilated @xmath6@xmath7 pair has @xmath128 , and the electron left over in the lowest ll has @xmath129 .
recombination of other , non - radiative states requires breaking the rotational symmetry ( e.g. , by interaction with other charges ) .
this result is independent of chosen spherical geometry ; on a plane the 2d translational symmetry leads to the conservation of both @xmath30 and @xmath31 , and the corresponding pl selection rule for @xmath85@xmath67 states is @xmath126@xcite .
three @xmath4 states in fig .
[ fig1 ] are of particular importance .
the @xmath8 and @xmath20 , the lowest singlet and triplet states at @xmath124 , are the only strongly bound radiative states , while @xmath15 has by far the lowest energy of all non - radiative ( @xmath130 ) states . in agreement with earlier prediction@xcite ,
the transition from @xmath8 to @xmath15 ground state is found at @xmath18 t , and a new , radiative excited triplet state @xmath20 is identified in all frames .
the binding energy @xmath104 of each of these three @xmath4 states , extrapolated to the @xmath131 ) limit , is plotted as a function of @xmath11 in fig .
[ fig2]a .
= 4.8 in clearly , the @xmath132 increases most quickly of all curves with increasing @xmath11 , the @xmath10 remains almost constant ( specially its coulomb part drawn with the thin line ) , and @xmath133 remains smaller than both @xmath132 and @xmath10 at any value of @xmath11 .
the extrapolated values of the oscillator strength @xmath89 of the @xmath3 and two radiative @xmath4 states are shown in fig .
the ratio @xmath134 remains almost independent of @xmath11 , and the resulting three pl peaks ( @xmath3 , @xmath8 , and @xmath20 ) are precisely the ones observed in experiments . in the above , we assumed that both electrons and holes are completely spin - polarized ( @xmath120 ) , but typically , all electron spins and only a fraction of hole spins @xmath135 are aligned with the field . as a result
, the @xmath20 pl has definite circular polarization ( @xmath136 ) and its intensity is reduced by @xmath135 , while the @xmath8 pl peak splits into a @xmath137 doublet ( separated by the appropriate zeeman energy ) with the intensity of the two transitions weighted by @xmath135 and @xmath138 .
an increase of @xmath135 from @xmath139 to 1 with increasing @xmath11 may explain the observed@xcite increase of @xmath140 by up to a factor of two , while @xmath141 remains constant .
[ [ interaction - of - charged - excitons - with - electrons . ] ] interaction of charged excitons with electrons .
even in dilute systems , recombination of bound @xmath6@xmath7 complexes can in principle be affected by their interaction with one another or with excess electrons .
specially the recombination of a @xmath15 might become allowed in a collision assisted process in which the translational symmetry of an isolated @xmath15 is broken .
the critical question is if the @xmath6@xmath4 correlations are of the laughlin type@xcite , meaning that the many - body @xmath6@xmath4 wave function contains a jastrow prefactor @xmath142 ( where @xmath143 and @xmath144 are complex coordinates of @xmath6 and @xmath4 , and @xmath145 is an integer ) .
if it is so , then a number @xmath145 of the highest energy @xmath6@xmath4 pair eigenstates are avoided in the low - energy many - body states @xcite ( just as the @xmath146 leading @xmath6@xmath6 pair eigenstates are avoided in the laughlin @xmath147 state of electrons ) .
this means lack of high - energy @xmath6@xmath4 collisions , and thus an effective isolation of the @xmath4 states from the 2deg and insensitivity of the @xmath4 binding or recombination to the electron density .
in particular , laughlin @xmath6@xmath15 correlations would eliminate the possibility of collision - assisted recombination of the @xmath15 . to determine
if the laughlin correlations occur in a mixed @xmath6@xmath4 liquid , one must calculate the @xmath6@xmath4 interaction pseudopotential @xmath148 , defined as the dependence of the pair interaction energy @xmath149 on the pair angular momentum @xmath50@xcite .
the general criterion for the occurrence of laughlin correlations in a many - body system confined to a degenerate ll and interacting through @xmath148 is that @xmath149 must have short range , i.e. decrease sufficiently quickly with increasing @xmath50@xcite . on a sphere
, @xmath149 must be a super - linear function of @xmath150 .
it turns out that this criterion is satisfied by the @xmath6@xmath4 repulsion in narrow qws@xcite .
this implies a simple connection between @xmath22 and the maximum allowed @xmath50 ( i.e. , minimum average separation ) for a @xmath6@xmath4 pair : @xmath151 for @xmath152 , and allows calculation of the effect of the @xmath6@xmath4 interaction on the @xmath4 recombination as a function of @xmath22 . in figs .
[ fig3 ] and [ fig4 ] we plot the pl oscillator strength @xmath89 and energy @xmath88 ( measured from the exciton energy @xmath119 ) for the @xmath153@xmath67 eigenstates corresponding to an @xmath4 interacting with another electron .
= 4.8 in = 4.9 in we assume that the zeeman energy will polarize all electron spins prior to recombination , except for those two in the @xmath8 , and concentrate on the following three initial configurations : @xmath6@xmath8 with @xmath154 and @xmath6@xmath20 and @xmath6@xmath15 with @xmath155 .
for each of the three configurations , @xmath89 and @xmath118 are plotted as a function of @xmath50 ( i.e. of @xmath22 ) .
for @xmath20 and @xmath15 , only an @xmath156@xmath157 pair can be annihilated , and an emitted photon has a definite circular polarization @xmath136 .
two indistinguishable electrons left in the final state have @xmath13 , so their @xmath50 must be odd ( @xmath158 minus an odd integer ) . for @xmath8 , both @xmath136 and @xmath159
pl are possible , with the energy of the latter transition shifted by @xmath115 . for @xmath136
, the two electrons in the final state can have either @xmath9 and @xmath50 even , or @xmath13 and @xmath50 odd ; while for @xmath136 they can only have @xmath13 and @xmath50 must be odd . as expected , for @xmath160 ( i.e. , @xmath161 ) both @xmath88 and @xmath89 converge to the values appropriate for single @xmath4 s plotted in fig . [ fig2 ] ( the energies shown in fig .
[ fig4 ] correspond to the @xmath136 transitions given by @xmath119 plus coulomb binding energy ; if present in the pl spectrum , additional @xmath159 transitions will appear at the energy higher by @xmath115 ) .
there is no significant effect of the @xmath6@xmath4 interactions on the @xmath4 recombination at small @xmath50 .
this justifies a simple picture of pl in dilute @xmath6@xmath7 plasmas , according to which , recombination occurs from a single isolated bound complex and hence is virtually insensitive@xcite to @xmath22 .
somewhat surprisingly , the laughlin correlations prevent increase of the @xmath15 oscillator strength through interaction with other charges ( @xmath162 remains ten times longer than @xmath163 even at @xmath24 ) .
this explains the absence of an @xmath15 peak even in the pl spectra @xcite showing strong recombination of a higher - energy triplet state @xmath20 ( except at very low temperatures@xcite ) .
an interesting feature in fig .
[ fig4 ] is the merging of @xmath164 and @xmath165 which actually has been observed@xcite .
the fractionally charged excitons ( fcx s ) @xmath7qe@xmath28 formed in strongly asymmetric qw structures ( @xmath5 ) consist of @xmath46 qe s bound to a valence hole @xmath7 . since qe s are collective excitations of a many - electron system , as many electrons as possible must be included in the computation .
this was only possible by a severe limitation of the single - particle hilbert space to the lowest ll .
although the inter - ll @xmath6@xmath7 scattering was quite important at @xmath79 ( where it caused binding of @xmath8 and @xmath20 states ) , it is greatly reduced at larger @xmath0 and , unlike at @xmath79 , the lowest - ll approximation is expected to be more justified . nevertheless ,
the model studied is necessarily a very ideal one and , consequently , such realistic elements as the finite qw width have also been excluded .
therefore , the numerical results obtained in this section are not meant to describe experiments as accurately as those for @xmath4 s , although the fact of the fcx binding or the optical selection rules for different fcx states are expected to be valid for experimental systems .
the binding of fcx states relies on the attractive interaction between the ( oppositely charged ) @xmath7 and qe , and the weak qe
qe interaction at short range@xcite . on the other hand , the stability of the fcx states at sufficiently large @xmath0 against the formation of @xmath3 or @xmath4 excitons results from the @xmath7qe attraction being sufficiently small compared to the laughlin gap @xmath166 .
the comparison of the @xmath7qe attraction energy ( the largest @xmath7qe pseudopotential parameter ) to @xmath167 shows that the fcx s rather than @xmath3 s or @xmath4 s should be the most stable bound states formed by a hole injected into the 2deg at @xmath0 larger than about a magnetic length @xmath1@xcite . [
[ numerical - energy - spectra . ] ] numerical energy spectra .
the series of @xmath168@xmath67 energy spectra calculated for @xmath169 and @xmath170 are presented in figs .
[ fig5 ] and [ fig6 ] .
= 4.9 in = 4.9 in the three spectra shown for each @xmath0 are obtained for different values of @xmath37 , corresponding to @xmath171 , 2 , and 3 qe s in the laughlin @xmath24 state of @xmath168 system ( without interaction with the valence hole ) . in fig .
[ fig5 ] ( @xmath169 , intermediate - coupling regime ) , new low - energy bands of states emerge in addition to those characteristic of strong coupling and containing an @xmath3 or @xmath4 .
these new states contain various fcx s interacting with the remaining qp s of the @xmath168 liquid .
in some cases the fcx states occur in the same spectrum with the @xmath3 or @xmath4 states .
for example , the @xmath7qe@xmath27qe band in fig . [ fig5]c coexists with the @xmath3 state ( meaning @xmath3 weakly coupled to the laughlin @xmath24 state of the remaining @xmath172 ) and the @xmath4qh band ( meaning @xmath15 laughlin - correlated with the remaining @xmath173 ; qh denotes a quasihole in the two - component , @xmath6@xmath4 laughlin liquid@xcite ) . in other cases ,
low - lying @xmath3 or @xmath4 states of the strong - coupling regime occur at the same @xmath50 as a low - lying fcx of the weak - coupling regime , and the transition between the two states ( which occurs at @xmath174 ) is continuous .
for example , @xmath7qe@xmath27 is mixed with @xmath4qh@xmath27 in fig .
[ fig5]b , and @xmath7qe is mixed with @xmath4qh@xmath175 in fig .
[ fig5]a . in fig .
[ fig6 ] ( @xmath176 , weak - coupling regime ) , well developed fcx bands occur .
the isolated @xmath7qe , @xmath7qe@xmath27 , and @xmath7qe@xmath175 states are the ground states in the spectra corresponding to @xmath171 , 2 , and 3 , respectively .
their angular momenta @xmath177 are obtained by adding @xmath178 and @xmath179 , where @xmath180 is the effective monopole strength in the composite fermion picture@xcite and @xmath181 .
similarly , the angular momenta of states containing an fcx and excess qp s result from adding @xmath177 and @xmath182 . [ [ selection - rules - and - photoluminescence . ] ] selection rules and photoluminescence .
similarly as it was for @xmath4 s , the translational symmetry of an isolated fcx leads to the conservation of @xmath50 and @xmath51 in the emission process .
this leads to the strict optical selection rules that can only be broken by collisions or disorder .
the recombination of an fcx state @xmath7qe@xmath28 formed in a laughlin @xmath147 electron liquid occurs through annihilation of a well defined number of qe s and creation of an appropriate number of qh s .
it turns out that the processes involving more than the minimum number of qp s all have negligible intensity , which for @xmath183 ( @xmath24 ) leaves only the following four possible recombination events : @xmath184 , where @xmath185 , 1 , 2 , or 3 , and @xmath186 denotes the photon .
let us apply the angular momentum conservation law , @xmath187 , to the above recombination events . in the fermionic picture@xcite ,
the angular momenta of qe and @xmath7 in the initial @xmath66@xmath67 state at a given monopole strength @xmath37 are @xmath188 and @xmath178 . by adding @xmath189 and @xmath190 , the following values are obtained for the angular momenta of fcx complexes@xcite : @xmath191 , @xmath192 , and @xmath193 .
the angular momentum of qh in the final @xmath194 state at the same @xmath37 is @xmath195 . by comparing the values of @xmath196 with
the angular momenta allowed for @xmath197 identical qh s in the final state , we obtain that ( i ) the `` decoupled hole '' state @xmath7 is radiative and can recombine to create a qh@xmath175 molecule with the maximum @xmath198 allowed for three qh s , ( ii ) @xmath199 is different from any @xmath50 allowed for two qh s and thus @xmath7qe is non - radiative ; however , the first excited state , @xmath7qe * at @xmath200 is radiative and recombines to create a qh@xmath27 molecule with @xmath201 ; ( iii ) @xmath202 is radiative and its recombination leaves behind a single qh ; ( iv ) neither @xmath203 nor its excitations are radiative .
the above analysis leaves @xmath7 , @xmath7qe * , and @xmath7qe@xmath27 as the only radiative fcx states , while @xmath7qe and @xmath7qe@xmath175 are found dark .
it is expected that a valence hole introduced into the 2deg at @xmath25 ( in the absence of free qe s ) and at @xmath5 will decouple and recombine from the initial state @xmath7 ( local filling factor @xmath24 ) . on the other hand , at @xmath26 the valence hole will bind one or more free qe s and recombine from either @xmath7qe * or @xmath7qe@xmath27 initial state , depending on @xmath0 , temperature , and the qe density . since the initial state from which the hole recombines changes at @xmath24 , and since different fcx states have different emission energy and intensity@xcite , the pl spectrum of a 2deg created in a strongly asymmetric structure is expected to change discontinuously at the corresponding magnetic field .
the pl spectrum of a 2deg was studied as a function of the separation @xmath0 between @xmath6 and @xmath7 layers .
two types of response of the 2deg to the optically injected hole were identified . in the strong - coupling regime ( @xmath204 )
the most strongly bound states are @xmath3 and @xmath4 , and the pl spectrum measures their optical properties rather than the original correlations of the 2deg .
in particular , the dark triplet @xmath15 remains virtually non - radiative at @xmath25 . in the weak - coupling regime ( @xmath5 ) , the @xmath4 states unbind and , depending on @xmath22 , the hole either decouples from the 2deg or binds one or more qe s to form a fcx state @xmath7qe@xmath28 .
the most stable radiative bound state at @xmath205 is found to be the `` decoupled hole , '' while at @xmath26 it is either @xmath7qe@xmath27 or @xmath7qe * ( excited state of @xmath7qe ) .
since different fcx states have different opticl properties , discontinuities are expected in the pl spectrum at @xmath24 . | the photoluminescence ( pl ) of a two - dimensional electron gas ( 2deg ) in a high magnetic field is studied as a function of the filling factor and the separation @xmath0 between the electron layer and the valence hole .
depending on the magnitude of @xmath0 relative to the magnetic length @xmath1 , two distinct regimes in the response of the 2deg to the valence hole occur , with different elementary emission processes contributing to the pl spectrum . at @xmath2 ( `` strong coupling '' regime ) ,
the hole binds one or two electrons to form an exciton ( @xmath3 ) or one of three possible charged exciton ( @xmath4 ) states , a spin - singlet or one of two spin - triplets . at @xmath5 ( `` weak coupling '' regime ) , the hole decouples or binds one or two laughlin quasielectrons to form fractionally charged excitons ( fcx s ) .
the binding energies as well as the emission energies and intensities of all @xmath4 and fcx states are calculated . |
recently , a remarkable connection between integrable models and combinatorics has emerged .
it first appeared in a series of papers concerning the xxz spin chain and the temperley
lieb ( tl ) loop model @xcite and which culminated with the so - called razumov
stroganov ( rs ) conjecture @xcite .
one of the main observations of @xcite , a weak corollary of the rs conjecture , is that the sum of entries of the properly normalized ground state vector of the tl(1 ) loop model is ( unexpectedly ! ) equal to the number of alternating sign matrices .
this result was eventually proved in @xcite by using the _ integrability _ of the tl loop model in the following way : the model is generalized by introducing @xmath2 complex numbers ( spectral parameters , or inhomogeneities ) in the problem , where @xmath2 is the size of the system .
the ground state entries become polynomials in these variables , and integrability provides many new tools for analyzing them , leading eventually to the exact computation of their sum , identified as the so - called izergin
korepin ( ik ) determinant , known to specialize to the number of alternating sign matrices in the homogeneous limit @xcite .
note that in this work , the meaning of the spectral parameters is not very transparent ; in particular , it is unclear how to generalize the full rs conjecture in their presence .
next , it was observed in @xcite that the polynomials obtained above really belong to a one - parameter family of solutions of a certain set of linear equations , in which the parameter @xmath0 has been set equal to a cubic root of unity .
this observation is not obvious because the equations for generic @xmath0 are not a simple eigenvector equation ; in fact , as explained in @xcite , they are precisely the quantum knizhnik zamolodchikov ( @xmath0kz ) equations at level 1 for the algebra @xmath3 . furthermore ,
in the `` rational '' limit @xmath4 , these polynomials have a remarkable geometric interpretation : they are equivariant hilbert polynomials ( or `` multidegrees '' ) of @xmath5 orbital varieties @xmath6 ( @xcite , see also @xcite ) , which are extensions of the joseph polynomials @xcite . note that here , the spectral parameters quite naturally appear as the basis of weights of @xmath7 . in @xcite , these ideas were generalized to higher algebras @xmath8 , which correspond to the orbital varieties @xmath9 . here , we pursue a _ different _ type of generalization : we investigate orbital varieties corresponding to the other infinite series of simple lie algebras : @xmath10 , @xmath11 , @xmath12 ; but we stick to the @xmath3 case by choosing the orbital varieties @xmath6 , @xmath13 a complex matrix in the fundamental representation .
indeed , we show below that such orbital varieties are related to the same loop model , but with different _ boundary conditions _ ( corresponding to variants of the temperley
lieb algebra ) .
furthermore , one can now @xmath0-deform the resulting polynomials to produce solutions of @xmath0kz equations of type @xmath14 , @xmath15 , @xmath16 and set @xmath0 to be a cubic root of unity .
taking the homogeneous limit , the entries become integer numbers , which we conjecture to be related to _ symmetry classes _ of alternating sign matrices and/or plane partitions ; in particular we identify the sums of entries . in what follows we state most results without proofs ; some will appear in a joint paper with a. knutson @xcite on a closely related subject .
let @xmath17 be a simple complex lie algebra of rank @xmath18 , @xmath19 a borel subalgebra .
@xmath20 where @xmath21 is the corresponding cartan subalgebra and @xmath22 is the space of nilpotent elements of @xmath19 .
@xmath14 and @xmath23 are borel and cartan subgroups .
let @xmath24 denote the weyl group of @xmath17 , and @xmath25 its standard generators , where @xmath26 runs over the set of simple roots of @xmath17 .
fixing an orbit @xmath27 , with @xmath28 and @xmath29 acting by conjugation , one can consider the irreducible components of @xmath30 , which are called orbital varieties .
even though much of what follows can be done for any orbital varieties , we focus below on the following special case : we fix an irreducible representation @xmath31 ( of dimension @xmath2 ) and consider the scheme @xmath32 .
the underlying set is precisely a @xmath30 , where @xmath33 is any element of @xmath34 such that @xmath35 is of maximal rank . in some sense ,
its components are the `` simplest possible '' orbital varieties .
it is known that there exists a representation of the weyl group @xmath24 on the vector space @xmath36 of formal linear combinations of orbital varieties ( springer / joseph representation ) ; for each @xmath29-orbit , it is an irreducible representation .
we use the following explicit form of the representation : note that orbital varieties are invariant under @xmath37 , where @xmath23 acts by conjugation and @xmath38 acts by overall scaling .
we can therefore consider equivariant cohomology @xmath39 and in particular via the inclusion map from each orbital variety @xmath40 to the space @xmath22 , the unit of @xmath41 is pushed forward to some cohomology class @xmath42 in @xmath43 $ ] , that is a polynomial in @xmath44 variables @xmath45 , @xmath46 , @xmath47 , @xmath48 ( the @xmath18 simple roots plus the @xmath38 weight ) , sometimes called _ multidegree _ of @xmath40 .
suppressing the @xmath38 action , that is setting @xmath49 , one recovers the joseph polynomials @xcite .
the way that @xmath24 acts on these polynomials can be described explicitly , by extending slightly the results of hotta @xcite to include the additional @xmath38 action .
one starts by associating to each simple root @xmath26 a certain geometric construction , which we briefly recall . for @xmath50 write @xmath51 where @xmath26 runs over positive roots , @xmath52 being a vector of weight @xmath26 .
define @xmath53 , and @xmath54 to be lvy subgroup whose lie algebra is @xmath55 . starting from an orbital variety @xmath40 , we distinguish two cases : * @xmath56 .
then set @xmath57 . * @xmath58 .
then let @xmath54 acts by conjugation : the top - dimensional components of @xmath59 are again orbital varieties ; set @xmath60 where @xmath61 is the multiplicity of @xmath62 in @xmath59 .
these elementary operations have a counterpart when acting on multidegrees , and a simple calculation shows that both cases are covered by a single formula : @xmath63 where @xmath64 is the reflection orthogonal to the root @xmath26 in @xmath65 $ ] , and @xmath66 is the associated _ divided difference operator _ , whereas on the left hand side @xmath25 implements right action on the @xmath42 , namely @xmath67 .
one can check that @xmath68 is a representation of the weyl group @xmath24 on polynomials .
note that at @xmath49 , we recover the natural action of @xmath24 ( up to a sign , with our conventions ) .
let us define the operator @xmath69 which acts in the space @xmath70 $ ] , @xmath71 being a formal parameter . rewriting slightly the relation above
we find that @xmath64 acts as @xmath72 .
using the fact that @xmath64 , just like the @xmath25 , satisfy the weyl group relations , we find that the operators @xmath73 also satisfy those . in the case of non - exceptional lie algebras , there are only 2 types of edges in the dynkin diagram , and therefore we have coxeter relations of the form @xmath74 , where @xmath75 depending on whether @xmath76 , there is no edge , a single or a double edge between @xmath26 and @xmath77 . writing these relations for @xmath78 and eliminating the @xmath64 , we find that relations with @xmath79 correspond respectively to the _ unitarity _ equation : @xmath80 the _ yang
equation : @xmath81 and the _ boundary yang baxter _ ( or reflection ) equation : @xmath82 whereas the case @xmath83 expresses a simple commutation relation for distant vertices .
indeed one recognizes in @xmath84 a standard form of the rational solution of the yang
baxter equation , the parameter @xmath71 playing the role of difference of spectral parameters . thus the multidegrees @xmath85 are closely connected to integrable models with rational dependence on spectral parameters , as will be discussed now .
before doing so , let us remark that in the special case investigated here of orbital varieties associated to @xmath6 , the @xmath25 obey more than just the coxeter relations .
in the @xmath86 case they actually generate a quotient of the symmetric group algebra @xmath87 known as the temperley
lieb algebra @xmath88 ( here 2 is the value of the parameter in the definition of the algebra , as will be explained below ) .
the same type of phenomena will be described for other simple lie algebras , and will lead to variants of the temperley
lieb algebra ; in particular , the `` bulk '' ( i.e. everything but a finite number of edges at the boundary ) of the dynkin diagrams being sequences of simple edges , these variants will only differ at the level of `` boundary conditions '' of the model .
let us now discuss the meaning of the equation @xmath89 where @xmath64 is the reflection associated to the root @xmath26 acting on the `` spectral parameters '' @xmath45 , @xmath46 , @xmath47 , @xmath72 is a certain linear operator defined above acting in the space @xmath70 $ ] and @xmath90 is a vector in that space .
when @xmath84 is the @xmath91-matrix ( or boundary @xmath91-matrix ) of some integrable model , such equations are satisfied by eigenvectors of the corresponding integrable transfer matrix .
more generally , these equations appear in the context of the quantum knizhnik zamolodchikov ( @xmath0kz ) equation , in connection with the representation theory of affine quantum groups @xcite . in either case
, it is known that we need an additional equation to fix the @xmath42 entirely .
define @xmath92 to be the semi - direct product of @xmath24 and of the weight lattice of @xmath17 .
it contains as a finite index subgroup the usual affine weyl group defined as the coxeter group of the affinized dynkin diagram . just like the affine weyl group
, it has a natural action on @xmath21 and therefore on @xmath65 $ ] which extends the action of @xmath24 generated by the reflections @xmath93 ; by definition , in this representation , an element of the weight lattice acts as translation in @xmath21 of the weight multiplied by @xmath94 ( @xmath95 where @xmath96 is the level of the @xmath0kz equation and @xmath97 is the dual coxeter number of @xmath98 ) .
then we claim that one can extend the representation of @xmath24 on @xmath70 $ ] ( the operators @xmath73 ) into a representation of @xmath92 , in such a way that each element of @xmath99 is the product of its natural action on @xmath65 $ ] and of a @xmath65$]-linear operator .
describing here the geometric procedure that leads to this action is beyond the scope of this paper .
the action will however be described explicitly in each of the cases below .
an important property is that if one sets @xmath49 the representation of @xmath92 factors through the projection @xmath100 .
so the @xmath38 action actually produces the affinization .
imposing that @xmath101 be invariant under the action of the whole group @xmath99 leads to a full set of equations , which are precisely equivalent to the so - called rational @xmath0kz equation ( or more precisely , a generalization of it for arbitrary dynkin diagram , the original @xmath0kz equation corresponding to the case @xmath86 ) at level @xmath102 ; and it turns out that they have a unique polynomial solution of the prescribed degree ( up to multiplication by a scalar ) .
the integrability suggests how to @xmath0-deform the above construction .
indeed , we have considered thus far @xmath91-matrices that form so - called _ rational _ solutions of the yang baxter equation , and @xmath101 s that are solutions of the rational @xmath0kz equation .
it is known however that the trigonometric @xmath91-matrices are a special degeneration of a one - parameter family of _ trigonometric _ solutions of the yang baxter equation , depending on a parameter @xmath0 .
setting @xmath103 , one customarily uses exponentiated multiplicative " spectral parameters of the form @xmath104 .
we then look for polynomial solutions @xmath101 of these parameters , to the corresponding trigonometric @xmath0kz equations .
the rational solutions are then recovered from the trigonometric ones via the limit @xmath105 , at the first non - trivial order in @xmath106 .
the details of the bulk and boundary @xmath91-matrices will be given below for the cases @xmath86 , @xmath10 , @xmath11 and @xmath12 .
we thus obtain , for any @xmath0 , a representation of the group @xmath99 , the @xmath24 relations satisfied by the @xmath73 and more generally the @xmath99 relations being undeformed . in terms of the new variables
@xmath104 living in @xmath23 , the natural action of an element of the weight lattice @xmath107 ( as the abelian subgroup of @xmath99 ) is the multiplication by @xmath108 . since for all simple lie algebras , @xmath107 has half - integer coordinates , we reach the important conclusion that when @xmath109 , this action becomes trivial . therefore , all operators associated to the weight lattice by the procedure outlined in the previous section become @xmath65$]-linear ( i.e. correspond to finite - dimensional operators on @xmath36 after evaluation of the parameters @xmath45 , @xmath46 , @xmath47 , @xmath48 ) . in this case
they are simply the scattering matrices of @xcite , and they commute with the usual ( inhomogeneous ) integrable transfer matrix of the model .
this implies that @xmath101 is an eigenvector of the latter ; in fact , we can call it `` ground state eigenvector '' because in the physical situation where the transfer matrix elements are positive , the perron frobenius theorem applies and the eigenvalue @xmath102 of @xmath101 is the largest eigenvalue in modulus . the value @xmath110 ( also called razumov stroganov point " ) is henceforth quite special and deserves a particular study . in particular , in the homogeneous limit where the spectral parameters @xmath111 are specialized to zero , @xmath101 can be normalized so that its entries are all _ non - negative integers _ , and we are interested in their combinatorial significance , in relation to the counting of alternating sign matrices and/or plane partitions .
we do not claim to have a full understanding of the general correspondence principle between simple lie algebras and these combinatorial problems , but we will perform a case - by - case study for @xmath86 , @xmath10 , @xmath11 and @xmath12 . a last remark is in order . as we shall see , it is simple to see that the solutions @xmath101 to the @xmath48 , @xmath14 , @xmath15 , @xmath16 @xmath0kz equations obey recursion relations , that allow to obtain the rank @xmath18 case from rank @xmath44 , hence we will content ourselves with the detailed description for @xmath18 with a given parity , namely @xmath112 , @xmath113 , @xmath114 , @xmath115 .
we review the @xmath86 case , already explored in @xcite .
we set @xmath116 , @xmath117 .
the fact that there are @xmath118 of these new variables @xmath119 , the spectral parameters , as opposed to the @xmath18 simple roots , is a reflection of the usual embedding @xmath120 .
@xmath19 ( resp .
@xmath22 ) is simply the space of upper triangular ( resp .
strictly upper triangular ) matrices of size @xmath2 , and the orbital varieties under consideration are the irreducible components of the scheme @xmath121 .
we also restrict ourselves to the case of @xmath122 even , which is technically simpler . in general , @xmath123 nilpotent orbits are classified by their jordan decomposition type , which can be expressed as a young diagram ; the orbital varieties are then indexed by standard young tableaux ( syt ) .
the condition @xmath6 ensures that only young diagrams with at most 2 rows can appear ( blocks in the jordan decomposition are of size at most 2 ) , and it is easy to check that all orbits are in the closure of the largest orbit , whose young diagram is of the form @xmath124 .
it is convenient to describe the corresponding syt by `` link patterns '' , that is @xmath2 points on a line connected in the upper - half plane via @xmath125 non - intersecting arches , see fig .
[ linkpatt ] .
the numbers in the first ( resp .
second ) row of the syt are the labels of the openings ( resp .
closings ) of the arches .
there are @xmath126 such configurations . 0.5 cm in this language , one has a rather convenient description of orbital varieties @xcite , which we mention for the sake of completeness . indeed , to each orbital variety @xmath40 we associate the upper triangular matrix @xmath127 with @xmath128 if points labelled @xmath129 and @xmath130 are connected by an arch , @xmath131 , @xmath132 otherwise .
then @xmath133 , @xmath14 acting by conjugation .
equivalently , @xmath40 is given by the following set of equations : ( i ) @xmath6 and ( ii ) @xmath134 , @xmath135 , where @xmath136 is the rank of the @xmath137 lower - left rectangle .
@xmath138 @xmath139 it is equally simple to describe the action of the weyl group , namely the symmetric group @xmath140 . rather than the generators corresponding to the simple roots : @xmath141 , @xmath117 used so far
, it proves simpler to consider the action of the projectors @xmath142 in the symmetric group algebra .
the operator @xmath143 acts on link patterns @xmath40 by connecting the arches ending at @xmath129 and @xmath144 and creates a new little arch between these 2 points ; this action is described on fig .
when a closed loop is formed , it is erased but contributes a weight @xmath145 .
the @xmath0-deformed version of this is obtained by attaching a weight @xmath146 to each erased loop , thus leading to the following ( pictorially clear ) relations : @xmath147=0\quad |i - j|>1\ ] ] all indices taking values in @xmath148 .
these are the defining relations of the temperley
lieb algebra @xmath149 .
when @xmath150 , i.e. @xmath145 , it is simply a quotient of the symmetric group algebra . alternatively , the deformed generators @xmath151 satisfy the usual relations of the hecke algebra ( of which the temperley
lieb algebra is a quotient ) .
in what follows , one special element of @xmath152 will be needed : it is the cyclic rotation @xmath153 .
its effect is to rotate the endpoints of the link patterns : @xmath154 without changing their connectivity .
it can also be expressed as : @xmath155 . for each simple root @xmath111 , we have the trigonometric @xmath91-matrix : @xmath156 where the @xmath157 generate @xmath152 and act in the space of link patterns as explained above .
we first write the system of equations : @xmath158 where @xmath159 acts by interchanging multiplicative spectral parameters @xmath160 and @xmath161 in the polynomial @xmath101 of the @xmath162 s , homogeneous of degree @xmath163 .
these equations are supplemented by the `` affinized '' equation satisfied by @xmath101 .
since the affine dynkin diagram @xmath164 is a circular chain , this equation quite naturally involves the cyclic rotation @xmath153 .
define the operator @xmath31 on @xmath165 $ ] which shifts the variables @xmath166 according to the rule : @xmath167 , @xmath168 and @xmath169 . then the additional equation is @xmath170 together with this equation , the above system forms the so - called level one @xmath0kz equation .
we claim that the @xmath171 and @xmath172 generate together @xmath99 . in order to see that , it is sufficient to build the @xmath2 generators @xmath173 of the abelian subgroup ( the lattice of weights ) .
they are given by @xmath174 , @xmath175 .
the original definition of the @xmath0kz equation is in fact the eigenvector equation for these `` scattering '' matrices ; with reasonable assumptions it is equivalent to the above system .
also , note that if one defines @xmath176 , then the @xmath177 , @xmath175 generate the usual affine weyl group ( a subgroup of order @xmath2 of @xmath99 ) .
the minimal degree polynomial solution of the level one @xmath0kz equation was obtained in @xcite , and is characterized by its `` base '' entry @xmath178 corresponding to the link pattern @xmath179 that connects points @xmath180 , with the value @xmath181 in which all factors are a direct consequence of the @xmath182 equations .
it is then easy to prove that all the other entries of @xmath101 may be obtained from @xmath178 in a triangular way .
_ example : _ at @xmath183 , there are 5 link patterns .
the minimal degree polynomial solution of the level one @xmath0kz equation reads : @xmath184 performing the rational limit @xmath185 , @xmath186 , @xmath103 yields the following multidegrees : @xmath187 and in particular the degrees @xmath188 respectively , upon taking @xmath189 and @xmath190 . at @xmath110 , @xmath101 becomes the ground state eigenvector of the integrable transfer matrix with periodic boundary conditions and inhomogeneities @xmath191 , @xmath46 , @xmath192 , or equivalently of the scattering matrices @xmath193 .
consider now the particular case @xmath194 , when @xmath101 is the perron
frobenius eigenvector of the hamiltonian @xmath195 where @xmath196 .
note that the periodic boundary conditions mean that @xmath197 is cyclic - invariant : @xmath198 . normalizing @xmath101 so that its smallest entry @xmath178 is @xmath102 , we have the following @xcite the sum of entries @xmath199 is equal to the number of alternating sign matrices , @xmath200 .
the result of @xcite is actually much more general , as the sum @xmath199 was evaluated in the presence of all the spectral parameters @xmath166 , and identified with proper normalization to the so - called izergin
korepin determinant @xcite , also equal to a particular schur function @xcite .
still unproven , however , is the @xcite the largest entry of @xmath101 , with arches connecting consecutive points , is @xmath201 .
for instance , plugging @xmath202 and @xmath110 into the above example , we get for @xmath183 , @xmath203 and @xmath204 , the total number of @xmath205 asms .
we now develop the @xmath10 case , which allows us to recover and interpret geometrically the results of @xcite .
we concentrate on the even case @xmath206 .
we parametrize as usual the roots @xmath116 for @xmath207 and @xmath208 .
we consider matrices that square to zero in the fundamental representation of dimension @xmath209 : a possible choice is to select upper triangular matrices satisfying @xmath210 , @xmath211 antidiagonal matrix with 1 s on the second diagonal .
it turns out that the orbital varieties are indexed by the same link patterns as before , of size @xmath18 ; and that the weyl group representation is actually a representation of the same quotient , the temperley
lieb algebra @xmath212 , the additional reflection @xmath213 being represented by a multiple of the identity .
according to the dicusssion above , the b @xmath0kz system reads : @xmath214 where @xmath215 stands for the inversion of the last spectral parameter , namely @xmath216 and @xmath217 is the degree of @xmath101 in @xmath218 .
finally , these equations are to be supplemented by the affinization relation . the latter is expressed by considering the reflection with respect to the extra root @xmath219 .
one finds that @xmath220 where @xmath221 is the degree of @xmath101 in @xmath219 . introducing the boundary operators
@xmath222 and @xmath223 so that eqs .
( [ firb][secb ] ) reduce to @xmath224 , as well as the usual @xmath225 , the generators of the weight lattice ( as abelian subgroup of @xmath99 ) are : ( i ) @xmath226 @xmath227 that implements @xmath228 and ( ii ) one additional generator implementing @xmath229 simultaneously for all @xmath129 .
the latter is a combination of * r * and * k * matrices as well as an additional operator implementing the reflection @xmath230 for all @xmath129 .
the minimal polynomial solution to the system ( [ firbx][secb ] ) has degree @xmath231 in each spectral parameter and total degree @xmath232 . as before it has a simple factorized base entry @xmath233 where @xmath234 is a common ( symmetric ) factor to all entries of @xmath101 .
all other entries may be obtained from this one in a triangular manner .
_ example : _ for @xmath235 , there are 2 link patterns as for the case @xmath236 .
the minimal degree polynomial solution of the level one @xmath235 @xmath0kz equation reads : @xmath237 as before , we get the corresponding multidegrees upon taking the rational limit , with the result : @xmath238 with @xmath239 ; hence the degrees @xmath240 for @xmath190 and @xmath189 . as explained in sect . 2
, the case @xmath110 is special in that the problem admits a transfer matrix , and its solution @xmath101 in the homogeneous limit where all @xmath202 is the groundstate of a hamiltonian @xmath241 which is the open boundary version of the @xmath86 hamiltonian @xmath197 . as shown in @xcite , at the rs point @xmath110 , and in the homogeneous limit where @xmath202 for all @xmath129 , and in which @xmath101 is normalized so that its smallest entry is @xmath242 , we have the following @xcite the sum of entries @xmath199 is equal to the number of vertically symmetric alternating sign matrices ( vsasm ) , @xmath243 .
this was actually proved in the same spirit as for the @xmath86 case , by identifying the sum of components including all spectral parameters @xmath166 as yet another determinant , which takes the form of a particular symplectic schur function .
a similar result holds for the case of odd @xmath244 , namely once properly normalized , the sum of entries @xmath199 is equal to an integer we call @xmath245 by analogy .
it turns out that @xmath246 is the number of cyclically symmetric transpose complement plane partitions ( cstcpp ) in an hexagon of size @xmath247 @xcite .
the numbers @xmath248 both have determinant formulae , namely @xmath249 , and @xmath250 . as in the @xmath48 case
, we have the @xcite the largest entry of @xmath101 , with arches connecting consecutive points , is @xmath251 .
example : for @xmath252 , taking @xmath253 and @xmath110 in the above expressions , we get the components @xmath254 , which sum to @xmath255 , the number of @xmath256 vsasms , and the maximal entry of @xmath101 is @xmath257 .
the simple roots of @xmath11 are @xmath116 , @xmath207 and @xmath258 .
we concentrate on the odd case @xmath259 , and consider the fundamental representation of dimension @xmath260 .
one choice is to select upper triangular matrices satisfying @xmath261 , @xmath211 antidiagonal matrix with @xmath102 s ( resp .
@xmath1 s ) in the upper ( resp .
lower ) triangle . there are @xmath262 orbital varieties , which are now indexed by _ open link patterns
_ , that is configurations of @xmath18 points on a line connected in the upper - half plane either in pairs via ( closed ) arches or to infinity via half - lines ( open arches ) .
the representation of the weyl group on these open link patterns takes the form of a modified temperley
lieb algebra .
we describe now its @xmath0-deformed version , @xmath263 ( see also @xcite for other variants of temperley lieb algebra ) . the generators @xmath264 obey the standard @xmath265 relations and the additional boundary " generator @xmath266 satisfies : @xmath267 , @xmath268 .
these generators act on open link patterns as follows .
open link patterns are represented with their open arches connected to a vertical line on the right .
the @xmath143 , @xmath269 act as usual , and @xmath266 like the left half of an @xmath270 , connecting the point @xmath271 to the vertical line ( first line of fig .
[ crules ] ) .
the rule is that any loop may be erased and replaced by a factor @xmath77 .
moreover , whenever a connection between points on the vertical line ( consecutive open arches ) is created , they may also be erased and replaced by a factor @xmath77 ( resp .
@xmath272 ) if this is created by the action of some @xmath273 ( resp .
@xmath274 ) . as @xmath18 is odd , the loop created by @xmath275 yields a weight @xmath77 , while that created by @xmath276 yields a weight @xmath272 , hence the result @xmath277 ( second line of fig .
[ crules ] ) .
we shall also need an additional operator @xmath278 satisfying the relations : @xmath279 and @xmath280 .
it is defined as @xmath281 , where @xmath282 is the involution acting on link patterns as follows : ( i ) @xmath283 if the arch connected to point @xmath102 is open , and ( ii ) @xmath284 otherwise , where @xmath62 is the link pattern in which the closed arch connected to @xmath102 is cut into two open arches .
to each simple root we attach respectively the standard trigonometric @xmath91-matrices @xmath285 , @xmath207 of eq . , and the boundary @xmath91-matrix @xmath286 , with the same expression . the level one @xmath15 @xmath0kz equation consists of the following system @xmath287 where as usual @xmath93 acts by interchanging the spectral parameters @xmath166 and @xmath161 , @xmath269 and @xmath215 acts on @xmath101 by letting @xmath288 , and @xmath217 is the degree of @xmath101 in @xmath218 .
these are finally supplemented by the affinization relation , obtained by considering an extra root , say @xmath289 , and the associated boundary operator @xmath290 : @xmath291 where @xmath292 interchanges @xmath293 and @xmath294 , and @xmath295 is of the form of eq . with @xmath296 in place of @xmath143 .
using @xmath297 , the relation can also be recast into @xmath298 the generators of the weight lattice ( as abelian subgroup of @xmath99 ) are very similar to the generators ( i ) of the case @xmath10 : the only change concerns the boundary operators @xmath222 and @xmath223 now implementing eqs . and .
the polynomial solution @xmath101 to the level one @xmath11 @xmath0kz system has degree @xmath299 in each variable , total degree @xmath300 and base entry @xmath301 and all the other entries of @xmath101 may be obtained in a triangular way from this one .
example : for @xmath302 , we have the following minimal polynomial solution to the level one @xmath303 @xmath0kz system : @xmath304 which , upon taking the rational limit yields the multidegrees : @xmath305 and the degrees @xmath306 for @xmath190 and @xmath189 . at the point @xmath110
, @xmath101 may be viewed as the ground state eigenvector of a transfer matrix , corresponding in the homogeneous limit to the hamiltonian @xmath307 normalizing @xmath101 so that its smallest entry @xmath242 , we have been able to compute the sum of entries to be @xmath308 . in the case of even @xmath206 , the above may be repeated almost identically : in the presence of spectral parameters , the even case may be recovered from the odd one by taking @xmath309 , and dividing out the result by @xmath310 . indeed , this specialization leaves us with only non - vanishing components whith an open arch at the rightmost point , in bijection with open link patterns with that point erased , hence the projection onto the case of size one less .
this leads us to the @xmath311 note that the sum in the even case , @xmath312 , also counts the cyclically symmetric self - complementary plane partitions ( csscpp ) in an hexagon of size @xmath247 @xcite .
also note the determinant formulae @xmath313 and @xmath314 .
furthermore , consider the left eigenvector @xmath315 of @xmath316 with the same eigenvalue ( @xmath18 for @xmath18 odd , @xmath317 for @xmath18 even ) .
normalize @xmath315 so that its entries are coprime positive integers .
we have found empirically the following @xmath318 finally , we formulate the the largest entry of @xmath101 for @xmath11 is the sum of entries for @xmath319 .
example : at @xmath320 , @xmath321 , @xmath322 , @xmath323 , @xmath324 , and the maximal entry of @xmath101 is @xmath325 .
the simple roots of @xmath12 are @xmath116 for @xmath326 and @xmath327 .
we concentrate on the odd case @xmath259 , and consider again the fundamental representation of dimension @xmath260 . just like in the @xmath10 case ,
one choice is to select upper triangular matrices satisfying @xmath261 , @xmath211 antidiagonal matrix with @xmath102 s on the second diagonal .
we now deal with @xmath16-type temperley
lieb algebras , denoted @xmath328 , with generators @xmath143 , @xmath269 obeying the @xmath265 relations and an extra generator @xmath329 , satisfying the relations : @xmath330 these operators act on open link patterns as follows .
the @xmath143 , @xmath207 act in the usual way , by creating a little arch between points @xmath129 and @xmath144 and by gluing the two former points . to describe the action of @xmath329 ,
let us first connect the open arches of the open link patterns by pairs of consecutive open arches from the left to the right , and represent the newly formed arches in a different color ( dashed lines , cf fig .
[ linkpad ] for the @xmath331 example ) .
we then define an involution @xmath282 on open link patterns that simply switches the color ( solid @xmath332 dashed ) of the rightmost arch if it is closed , and leaves it invariant if it is open
. then @xmath333 .
finally , we introduce an extra boundary operator @xmath334 , which is the right half of an @xmath270 ( like a reflected @xmath266 of @xmath11 ) , with its open end connected to the vertical line , and acts as such , with the same rules as for @xmath11 , but upon reflection of indices @xmath335 .
it satisfies the relations : @xmath336 and @xmath337 .
the level one @xmath16 @xmath0kz equation consists of the following system @xmath341 where as usual @xmath93 acts by interchanging the spectral parameters @xmath166 and @xmath161 , @xmath269 and @xmath342 acts on @xmath101 by interchanging @xmath343 and @xmath344 . upon using the above relation @xmath333 , the latter equation may be equivalently replaced by @xmath345 these are finally supplemented by the affinization relation , obtained by considering the extra root @xmath346 , and the associated boundary operator @xmath347 involving the extra operator @xmath334 : @xmath348 where @xmath349 and @xmath221 the degree of @xmath101 in @xmath191 .
the minimal degree polynomial solution to the level one @xmath12 @xmath0kz system has total degree @xmath350 and partial degree @xmath351 in all variables .
its base entry , corresponding to the open link pattern @xmath179 with only open arches reads @xmath301 and all the other entries of @xmath101 may be obtained in a triangular way from this one .
example : for @xmath302 , we have the following minimal polynomial solution to the level one @xmath352 @xmath0kz system : @xmath353 which , upon taking the rational limit gives the multidegrees : @xmath354 and the degrees @xmath355 for @xmath190 and @xmath189 . at the point @xmath110
, @xmath101 may be viewed as the perron
frobenius eigenvector of a transfer matrix , corresponding in the homogeneous limit to the hamiltonian @xmath356 note that upon the reflection @xmath357 , this hamiltonian is mapped onto @xmath316 : we are dealing with the same algebra , but in different representations .
this conjecture also works in the even case @xmath206 , which may be obtained from the odd one by taking @xmath359 , shifting all remaining spectral parameters @xmath360 , @xmath361 , and dividing out by @xmath362 .
note the formulae @xmath363 and @xmath364 .
razumov and yu.g .
stroganov , spin chains and combinatorics , _ j. phys _ a34 ( 2001 ) , 3185 , cond - mat/0012141 ; spin chains and combinatorics : twisted boundary conditions , _ j. phys _ a34 ( 2001 ) , 5335 , cond - mat/0012247 | we study a one - parameter family of vector - valued polynomials associated to each simple lie algebra .
when this parameter @xmath0 equals @xmath1 one recovers joseph polynomials , whereas at @xmath0 cubic root of unity one obtains ground state eigenvectors of some integrable models with boundary conditions depending on the lie algebra ; in particular , we find that the sum of its entries is related to numbers of alternating sign matrices and/or plane partitions in various symmetry classes . |
during the past decade the field of molecular magnets has experienced a rapid evolution.@xcite nowadays a vast variety of species can be synthesized , ranging in size from 2 to more than 30 paramagnetic ions embedded in the host molecule.@xcite the fascinating properties of these new materials include hysteretic behavior,@xcite quantum tunneling of the magnetization,@xcite magnetocaloric,@xcite and magnetostrictive effects.@xcite astonishingly , for several observables and not too low temperatures it is sufficient to treat magnetic molecules classically by applying the heisenberg model.@xcite this enables one to apply the powerful machinery of classical stochastic sampling methods .
the metropolis algorithm @xcite has become the standard tool to calculate statistical classical properties of these nano magnets.@xcite recently , the wang - landau algorithm@xcite ( wl ) has been successfully applied to various problems in statistical physics and biophysics both to models with discrete @xcite as well as with continuous degrees of freedom.@xcite in this context the wl exhibits a superior feature in comparison to the metropolis algorithm . since the wl calculates the density of states ( dos ) , one can estimate thermodynamic observables such as the free energy and entropy for all temperatures using just one single simulation of the density of states .
however , the calculation of the dos in energy space @xmath4 permits only to calculate thermodynamic properties as a function of temperature and at zero magnetic field , e.g. the zero - field specific heat , but not observables such as the magnetic susceptibility at non - vanishing field.@xcite for studying the magnetic properties of a system at arbitrary external magnetic field one has to calculate a joined dos @xmath3 in energy and magnetization space .
once @xmath3 is known , properties like magnetization @xmath5 and magnetic susceptibility @xmath6 can be calculated at any temperature and any external magnetic field using again just one single simulation of the density of states.@xcite in this article we introduce a self - adaptive version of the wl which allows to calculate the two - dimensional dos @xmath3 using a discrete binning scheme in the continuous energy and magnetization space . we demonstrate that the wl is capable to calculate efficiently all thermodynamic properties of rather large spin systems using the example of the magnetic keplerate molecule \{mo@xmath0fe@xmath1}.@xcite the dependencies of the thermodynamic observables both on temperature and external magnetic field can be obtained by only one single simulation .
we conclude by discussing the problems arising at low temperatures and high magnetic fields .
in the keplerate molecule \{mo@xmath0fe@xmath1 } 30 iron ( iii ) ions ( @xmath7 ) occupy the vertices of a perfect icosidodecahedron,@xcite see fig .
[ f - a ] .
the spins are coupled with their nearest neighbors by an isotropic and antiferromagnetic coupling of strength @xmath8
k. the spectroscopic splitting factor is @xmath9.@xcite fe@xmath1 } the vertices represent spin sites and the edges represent nearest neighbor interactions.,width=151 ] we write the heisenberg hamiltonian as @xmath10 whereas @xmath11 directs that the sum is over distinct nearest - neighbor pairs , @xmath12 is an external magnetic field in @xmath13-direction , @xmath14 is the bohr magneton , and @xmath15 denotes classical vectors of length @xmath16.@xcite we use a self - adaptive @xcite scheme of the wl to calculate the two - dimensional dos @xmath3 , where @xmath17 denotes the heisenberg energy of a given spin configuration of hamiltonian ( [ hamiltonian ] ) without external magnetic field , i.e. @xmath18 and @xmath5 is the magnetization in @xmath13-direction , which is defined as the sum over the @xmath13-components of all classical spin vectors : @xmath19 in contrast to the metropolis algorithm , where the acceptance probability of a new generated state is determined by @xmath20,1\}$ ] , the wl is characterized by an acceptance ratio @xmath21 , where @xmath22 and @xmath23 refer to energies before and after the transition .
the original wl has been applied to models where the energy assumes only discrete values.@xcite since the heisenberg model consists of classical spin vectors which are continuously orientable , the possible energy and magnetization values are real numbers between the minimum and maximum energy and magnetization , respectively .
thus , we discretize the continuous energy and magnetization range by the introduction of bins.@xcite we have chosen bins of uniform width of @xmath24 k in energy and @xmath25 in magnetization range . to speed up the simulation we divide the total energy range in eight overlapping intervals and follow the recipe in ref . to avoid boundary effects
a priori it is not known , which magnetization bins are accessible by the random walker in a certain energy interval . a striking feature of the wl is , that one does not have to know anything about the dos one wants to calculate .
following the original wl , the random walker biases itself to explore the accessible energy and magnetization space .
to do so , we have chosen the following procedure : at the beginning we perform a trial run for the desired energy interval .
we introduce an initial dos @xmath26 and a reference histogram @xmath27 , and perform spin flips according the original wl procedure accepting and neglecting new spin configurations for a defined number of monte - carlo steps @xmath28 .
one monte - carlo step corresponds to a single spin tilt event . each time
a bin ( @xmath29 ) is visited the corresponding entries in @xmath30 and @xmath31 are updated , whereas @xmath32 is the initial modification factor . following this recipe
the random walker triggers itself to explore the accessible energy and magnetization space for a predefined energy interval . for the molecule \{mo@xmath0fe@xmath1}.
@xmath33 defines the accessible energy and magnetization space .
the grid shows the bins in energy and magnetization space.,width=226 ] after @xmath28 steps have been performed we stop the initial run .
now we continue with the normal wl for the same energy interval . as an initial guess for the dos we use @xmath34 .
a new state is only accepted if the corresponding entry in @xmath27 is not zero , compare fig .
[ f - b ] .
in other words , we accept only states corresponding to bins in the energy and magnetization space which have already been visited during the initial run , otherwise the new generated state is neglected according to ref . .
in addition , the relative flatness of the accumulated histogram is only checked after all valid entries in the reference histogram have been visited at least one hundred times .
of course the total runtime of the simulation as well as the accuracy of the finally calculated dos are very sensitive to a good choice of @xmath28 .
if @xmath28 is chosen very large , the random walker explores bins at the boundaries of the accessible energy and magnetization space , which are rarely visited .
thus the dos for these entries is small and does not contribute significantly to the partition function .
nevertheless , sampling of these states increases the total runtime of the simulation since a flat histogram is to be accumulated during the simulation . on the other hand ,
if @xmath28 is chosen too small , the energy and magnetization space is sparsely explored and many bins of the accessible energy and magnetization space are not visited , resulting in an inaccurate estimate of the dos .
thus , the question is to find a good tradeoff between runtime and accuracy of @xmath3 . to estimate a reasonable value for @xmath28 for each energy interval
, we have chosen different values and counted the number of visited distinct bins in @xmath27 after finishing the initial run .
it turns out that the number of visited bins shows some saturation behavior . after a critical number of steps
the visited energy and magnetization space does not grow significantly any more .
thus , after this critical number of steps has been reached we consider the accumulated @xmath27 to be a good estimate for the successive wl run . as the initial modification factor @xmath32 we have chosen a rather large value of @xmath35 and
reduce @xmath32 in large steps.@xcite after finishing the initial run we perform 14 steps decreasing @xmath32 according to the recipe @xmath36{f_i}$ ] resulting in a final modification factor of 1.000000015 .
after @xmath3 has been obtained , the magnetization @xmath37 as a function of temperature @xmath38 and external magnetic field @xmath39 can be calculated from @xmath40 the differential magnetic susceptibility @xmath41 can be computed by using the equation @xmath42.\ ] ] while without external magnetic field it is sufficient to calculate only the dos for negative energies , since states with positive energies due to their boltzmann factor @xmath43 practically do not contribute to the partition function , in the case of an external magnetic field also energy bins with a positive energy become important , since they might be low - lying in the case of an applied external field .
thus it is required to calculate the dos over the full energy range of the system , if the properties at high magnetic fields are studied . the exact ground state energy for \{mo@xmath0fe@xmath1 }
is known to be @xmath44 k,@xcite which is close to the lowest accessible energy bin of @xmath45 k. the classical ground state of the molecule is charactereized by relative angles of 120@xmath46 between nearest - neighbor spins.@xcite the possibility of generating such a spin configuration is practically zero , and thus an effective sampling of the dos near the ground state energy is not possible .
the same is valid for the highest energy states of the molecule , where the probability of generating a spin configuration with all spins pointing in the same direction is unlikely .
the highest energy bin visited during the initial run is @xmath47 k. as the criterion of flatness we used a relatively small value of 0.5 , and to further decrease the statistical error we perform four runs , calculate the magnetic properties for each of these runs and compute the mean value and mean standard deviation afterwards . on a personal computer ( intel , xeon , 2.60 ghz ) the sampling of the complete dos took about 120 hours of cpu time . for a comparison of accuracy we performed extensive monte - carlo simulations using the metropolis algorithm .
we perform @xmath48 spin tilt trials to let the system reach equilibrium at a defined external magnetic field and temperature . afterwards for additional @xmath49 tilting trials the thermodynamic properties are computed .
( color online ) joint dos @xmath50 $ ] for the magnetic molecule \{mo@xmath0fe@xmath1}. ] ( color online ) joint dos @xmath50 $ ] for the magnetic molecule \{mo@xmath0fe@xmath1}. ] the calculated joint dos @xmath3 for the magnetic molecule \{mo@xmath0fe@xmath1 } is presented in fig . [ 2d_dos ] .
one notices that the low - energy boundary @xmath51 assumes a parabolic shape as observed for various systems.@xcite the region with @xmath52 at high energies indicates the global rotational symmetry of the system .
one also notices that in the vicinity of the phase - space boundary the classical density of states grows very rapidly by orders of magnitude which explains why it is difficult to obtain very accurate results at the boundaries .
( color online ) magnetization @xmath5 as a function of temperature @xmath38 and external magnetic field @xmath39 . ]
the limited accuracy at the energy and magnetization space boundaries looses its significance with increasing temperature .
this is demonstrated by evaluating the magnetization as a function of temperature @xmath38 and external magnetic field @xmath39 .
figure [ mvstb ] shows the behavior of the magnetization in the relevant temperature and field range .
the simulated magnetization does not show any significant statistical fluctuations ; it compares nicely to the result of a metropolis sampling .
( color online ) comparison of the magnetization @xmath5 as a function of external magnetic field calculated from @xmath3 ( solid lines ) with computed data using the metropolis algorithm at @xmath53 k ( open circles ) and 4 k ( open squares ) , respectively . statistical sampling errors are smaller than the used symbols and line widths . ]
figure [ m_wl_metr_theory ] compares the magnetization at @xmath54 k and 4 k as a function of field for the wl and the metropolis simulations . as can be inferred from the figure , at a temperature of the order of the exchange coupling the wl reaches the same accuracy as the metropolis algorithm , but the latter has to be performed for each pair of variables @xmath55 , whereas with the wl the density of states has to be sampled only once in order to obtain the complete function @xmath56 .
( color online ) magnetic susceptibility @xmath6 as a function of temperature @xmath38 and external magnetic field @xmath39 . ] ( color online ) comparison of the magnetic susceptibility @xmath6 as a function of external magnetic field calculated from @xmath3 ( solid lines ) with computed data using the metropolis algorithm at @xmath53 k ( open circles ) and 4 k ( open squares ) , respectively . ] the differential susceptibility , figs .
[ suszvstb ] and [ suszvsb ] , is a second derivative of the partition function .
therefore , it will magnify inaccuracies of the simulated density of states .
figure [ suszvstb ] displays the susceptibility as a function of @xmath38 and @xmath39 .
it can be seen that for temperatures higher than @xmath57 the behavior is rather smooth , but for lower temperatures a spiky behavior is observed which results from statistical fluctuations at the energy and magnetization space boundary . at @xmath58
these fluctuations are still visible , but on average much smaller than the real minimum at about @xmath59,@xcite compare fig .
[ suszvsb ] .
clearly one can see that the dip in suszeptibility at about @xmath59 vanishes with increasing temperature.@xcite the spurious peak at @xmath60 and @xmath61 reflects the difficulty to obtain the density of states in close vicinity of the ground state . a comparison with metropolis simulations is shown is fig .
[ suszvsb ] .
the results obtained using both methods agree well .
the fluctuations in the wl data are due to the missing bins in @xmath3 at the energy and magnetization space boundaries , since these features are apparent in all four wl runs .
the statistical sampling errors are given by the size of the symbols and by the line widths used in fig .
[ suszvsb ] .
( color online ) specific heat @xmath62 as a function of temperature @xmath38 and external magnetic field @xmath39 . ]
the last function we want to discuss in this paper is the specific heat @xmath62 as a function of temperature @xmath38 and external magnetic field @xmath39 . as can be seen in fig .
[ shvstb ] the specific heat does not show too strong fluctuations at low temperature and fields below saturation .
it seems that this observable is more robust against statistical fluctuations of the density of states than the susceptibility . in order to complete the discussion we display the specific heat also for magnetic fields above the saturation field where it shows strong fluctuations .
they reflect magnified inaccuracies at the high energy boundary of @xmath3 and thus are spurious . summarizing , one can say that the proposed self - adaptive version of the wang - landau algorithm can efficiently generate the density of states @xmath3 of a rather large spin system , such as the magnetic molecule \{mo@xmath0fe@xmath1}. the obvious advantage is that with one simulation of @xmath3 many thermodynamic observables can be evaluated as function of temperature and applied magnetic field .
statistical fluctuations are apparent in the second derivatives of the two - dimensional dos , namely the specific heat and magnetic suszeptibility at low temperatures .
these fluctuations can be minimized at the cost of computational time , by either increasing the number of bins or by a more strict flatness criterion .
recently , it was demonstrated that a fitting of the one - dimensional dos for a sampled energy interval by higher order polynomials results in a less fluctuating specific heat.@xcite nevertheless , it is stated , that this fitting procedure fails if one tries to extrapolate the dos outside the sampled energy interval using the fitting function .
thus , further improvements are still needed to provide an effective sampling at the boundaries of @xmath3 . due to the fact that the density of states varies by many orders of magnitude it is clear that close to the boundaries of @xmath3 the statistics becomes poorer .
the authors are indebted to d.p .
landau for having initiated this work .
this work was supported by the deutsche forschungsgemeinschaft ( grant no .
schn 615/5 - 1 ) . | the wang - landau method is used to study the magnetic properties of the giant paramagnetic molecule \{mo@xmath0fe@xmath1 } in which 30 fe@xmath2 ions are coupled via antiferromagnetic exchange .
the two - dimensional density of states @xmath3 in energy and magnetization space is calculated using a self - adaptive version of the wang - landau method . from @xmath3
the magnetization and magnetic susceptibility can be calculated for any temperature and external field . |
we start our experiments by preparing two @xmath17li atoms in the @xmath46 and @xmath47 hyperfine states labeled @xmath0 and @xmath1 in the motional ground state of an optical microtrap . to do this
we follow the procedure established in serwane et al.@xcite .
we first prepare a two - component fermi gas consisting of @xmath48 atoms in a crossed - beam optical dipole trap at a temperature of @xmath49nk .
this dipole trap acts as a reservoir from which we load about 1000 atoms into our microtrap , which is created by focusing a far red - detuned laser beam ( @xmath50 nm ) with a custom - designed high - resolution objective ( @xmath51)@xcite .
we then apply a magnetic field gradient to deform the trapping potential until only a single bound state remains in the microtrap .
this spilling technique allows us to prepare systems of one @xmath0 and one @xmath1 atom in the ground state of the trap with a typical fidelity above @xmath52 . to characterize our microtrap ,
we have measured its trapping frequencies using modulation spectroscopy .
we find trap frequencies of @xmath53khz and @xmath54khz at an optical power of @xmath55w , which is in good agreement with our expectations for a cigar - shaped dipole trap with a waist of @xmath56 and an aspect ratio @xmath57 .
to form a double - well potential , we create two microtraps which are partially overlapping .
we achieve this by using an acousto - optical deflector ( aod ) to create two separated laser beams that are incident upon the objective under slightly different angles , which the objective translates into a spatial separation of the two foci in the focal plane . to control the spacing and the tilt of the resulting double - well potential we tune the relative angle and intensity of the two trapping beams .
we do this by changing the frequencies and amplitudes of two rf - signals driving the aod . to obtain a tunnel coupling which is on the order of 100hz at our typical light powers of @xmath58w
, we set the frequencies to values of 32mhz and 38mhz which corresponds to a spacing of the double well of @xmath59 m . to tune the strength of the tunnel coupling during our experiments , we change the overall depth of the potential which we control by actively stabilizing the overall light power @xmath60 of the trapping beams . at these trap parameters ,
the individual wells have a depth of approximately @xmath61{khz}$ ] and the height of the potential barrier between them is approximately @xmath62{khz}$ ] . to compare our double well to optical lattice experiments
, we can define an effective lattice wavelength @xmath63 and calculate the recoil energy @xmath64{khz}$ ] , where @xmath65 is the planck constant and @xmath66 is the mass of @xmath17li .
since our experiment allows us to independently address the individual wells of our potential , we can perform site - selective atom - number detection by switching off one of the wells and measuring the number of atoms in the remaining well . to do this
, we first decouple the wells by rapidly increasing the height of the potential barrier within 2ms .
this ramp is fast compared to the time scale of the tunnel coupling but slow enough to avoid heating of the atoms into higher trap levels .
then we slowly turn off one of the wells while keeping the overall light power constant . during this process
, the atoms in this well are adiabatically transferred into excited states of the remaining well .
therefore , we use the spilling technique introduced above to remove all atoms from these highly excited states . to count the number of atoms in the remaining well , we release the atoms from the microtrap , recapture them in a magneto - optical trap and measure their fluorescence with a ccd camera ( andor ixon dv887 ) for an exposure time of 1 second @xcite .
to experimentally realize the hubbard model , our double - well system has to fulfill the tight - binding approximation and the single - band approximation . in a lattice potential
, the wavefunction of a single particle can be described as a superposition of localized wavefunctions centered on the individual sites of the lattice . in the limit of tight binding , the overlap between wavefunctions located on different sites is small and they can be approximated by the eigenstates of the individual wells . in this case , the tunnel coupling @xmath10 between adjacent sites is much smaller than the lowest excitation energy on the individual wells . since in our double well ,
the tunneling constant @xmath10 is approximately @xmath67{hz}$ ] and the axial trap frequency @xmath68 in each well is about @xmath69{khz}$ ] , the tight - binding approximation is well fulfilled in our system . in the single - band approximation
, one only considers the motion of the particles in the lowest bloch band , which is composed of the ground states of the individual wells . to fulfill this approximation , all energy scales in the system have to be much smaller than the band gap to the first excited band . in our double - well system , we initialize both atoms in the ground state of one well ( see sec.[sec : prep ] ) and limit the interaction energy @xmath11 to values smaller than the lowest excitation energy of the single wells during all our experiments . for all but the largest values of the interaction energy we estimate the wavefunction overlap between two interacting particles localized in a single potential well and the respective non - interacting ground state to be above @xmath70{\%}$ ] . for the largest interaction energies of @xmath71{hz}$ ] for the dataset with @xmath72{hz}$ ] ( @xmath73{hz}$ ] for @xmath74{hz}$ ] ) the overlap decreases to @xmath75{\%}$ ] ( @xmath76{\%}$ ] ) . using these two approximations ,
the spatial wavefunction of a single particle in our double - well system can be described as a superposition of the ground - state wavefunctions of a particle in the left @xmath77 or the right @xmath78 well .
this results in a two - level system where the states @xmath77 and @xmath78 are coupled with a tunnel coupling @xmath10 .
we calibrate the hubbard parameters of our double well by observing the tunneling dynamics in our system . to determine @xmath10 and @xmath27 , we study the tunneling of two non - interacting particles between the wells ( sect.[sec : tunneling ] and sect.[sec : tilt ] ) .
we then determine the on - site interaction energy @xmath11 for different strengths of the interparticle interaction ( sect.[sec : onsite ] ) . to prepare the initial state for our tunneling experiments , we start from two non - interacting atoms in the ground state of a single microtrap ( see sec.[sec : prep ] )
. then we slowly ramp on a second trapping beam ( see sec.[sec : dwpotential ] ) and thereby change the shape of the potential from a single gaussian well to a double - well potential . during this ramp ,
we linearly increase the overall power @xmath79 to keep the depth of the initial well constant . at any time in this process
, @xmath79 is large enough to prevent the atoms from tunneling between the wells . to start the tunneling dynamics , we lower the height of the barrier between the two wells within @xmath80ms . since this ramp
is fast compared to the time scale of the tunnel coupling we can treat it as a sudden switch - on of the tunneling . yet , this ramp is slow enough to avoid exciting the atoms into higher bands .
we then let the system evolve for different hold times , ramp up the barrier to decouple the wells and finally measure the number of atoms in one of the wells , using the detection scheme described above .
as an example , the tunneling oscillation of two non - interacting atoms in a symmetric double - well at an overall power of @xmath81{\mu w}$ ] is shown in fig.1 . to determine the value of the tunnel coupling @xmath10 we fit the observed oscillation with a damped sine wave and find @xmath82hz and a damping time of @xmath83ms .
we believe that the main cause for the observed damping is dephasing due to long - term drifts of the tilt of the double - well potential .
to calibrate @xmath10 for different barrier heights , we perform these measurements for different values of the overall light power @xmath79 ( fig.2a ) . for our experiments , we use light powers of @xmath84w and @xmath85w which result in tunnel couplings of @xmath86hz and @xmath87hz .
tunneling between two states that are separated by a potential barrier is only possible if their energy difference @xmath27 is not much larger than their tunnel coupling @xmath10 . to achieve this for our double - well system
, we control the relative depth of the two wells by changing the relative power of the rf signals driving the aod ( see sec.[sec : dwpotential ] ) .
to calibrate the tilt @xmath27 , we measure the oscillation frequency for various values of the relative rf power ( fig.2b ) .
we fit the measured frequencies with the effective coupling @xmath88 expected for a two - level system .
we find good agreement to the data and therefore conclude that our non - interacting system can be described by the two - site hubbard model ( see sec.[sec : hubbard ] ) .
using this fit , we calibrate the tilt @xmath27 in units of the tunnel coupling @xmath10 and extract the relative rf power where the double well is balanced ( @xmath33 ) .
to observe coherent tunneling in our double - well system , we have to stabilize the relative depth of the two wells with a precision better than @xmath10 . for our typical trap depth of @xmath89{khz}$ ] and
tunnel couplings of @xmath90{hz}$ ] ( at an overall light power of @xmath91{\mu w}$ ] ) , this corresponds to a relative stability of the intensities of the two beams that is on the order of @xmath92 .
however , switching on the second well leads to heating of the aod which results in a slow drift of @xmath27 on the order of @xmath93{hz}$ ] over several seconds .
since these drifts are the same in every experimental cycle , we can compensate them by applying an exponential ramp to the relative rf power . using this technique ,
our stability is currently limited by slow drifts of the tilt which are on the order of @xmath94{hz}$ ] to @xmath95{hz}$ ] per day .
to control the on - site interaction energy @xmath11 between two particles that occupy the same well , we change the s - wave scattering length @xmath96 between the atoms with a magnetic feshbach resonance . to calibrate the interaction energy as a function of the magnetic offset field , we study the tunneling of two interacting particles . for a system with two particles in one well
, the energy is shifted with respect to the non - interacting system by the interaction energy @xmath11 . if the two particles tunnel together , this interaction energy is conserved and consequently pair tunneling is resonant in a symmetric double well at all interaction strengths . however
, if a single particle tunnels to the other well the particles can no longer interact and hence single - particle tunneling is effectively detuned by the interaction energy @xmath11 .
we can compensate this effective detuning by a tilt of @xmath97 between the two wells@xcite .
this allows us to determine the on - site interaction energy by measuring the value of the tilt at which single - particle tunneling is restored . to perform this calibration
, we measure the amplitudes of single - particle and pair tunneling as a function of tilt .
to do this we initialize the system with two interacting atoms in the left well , let them tunnel between the wells for different hold times and measure the number of atoms in the right well . from these measurements ,
we extract the probabilities of finding 1 or 2 atoms in the right well , average over all hold times and plot them as a function of the tilt ( fig.2 of main text ) .
we observe that the tilt at which single - particle tunneling is resonant shifts with the interaction strength whereas the pair - tunneling resonance remains at @xmath33 . by fitting lorentzians to the resonance peaks and measuring their separation , we can therefore determine the value of the on - site interaction @xmath11 for different magnetic fields ( fig.3a ) . since the value of the on - site interaction energy depends on the depth of the confining potential
, we perform the calibration of @xmath11 for the two potential depths ( see sec.[sec : onsite ] ) we use in our experiments . to check whether the measured interaction energies are affected by off - site interactions we compare our results to a calculation of the interaction energy of two particles in a cigar - shaped potential ( fig.3b)@xcite . within the errors we find good agreement with our data and
therefore conclude that the effects of off - site interactions in our experiments are small .
the good agreement between the measured and calculated interaction energies allows us to use the results from this calculation as the calibration for @xmath11 for all our measurements .
to prepare two atoms in an eigenstate of the symmetric double well , we start with two non - interacting atoms in the ground state of a single well and adiabatically change the shape of the potential to a double well . to prepare the system in the ground state @xmath98 , we initialize the system in state @xmath99 , but keep the left well much deeper than the right one .
then , we lower the potential barrier between the wells and thereby set the tunnel coupling @xmath10 to a value of about @xmath67{hz}$ ] . by setting a tilt of @xmath100
, we make sure that tunneling is far off - resonant ( @xmath101 ) and the initial state @xmath99 has almost complete overlap with the ground state of the tilted double well ( fig.4 ) . by adiabatically reducing the tilt to @xmath102 within 100ms we then bring the system into the ground state of the symmetric double well . using the same technique
, we can also prepare the atoms in the highest energy state @xmath44 . to do this , we start again in state @xmath99 , but this time we set the right well to be much deeper than the left one ( @xmath103 ) . lowering the barrier
then transforms the initial state @xmath99 into state @xmath44 ( fig.4 ) . after preparing the non - interacting eigenstate of the double - well system
, we adiabatically introduce interparticle interactions by ramping the magnetic field from the zero - crossing of the scattering length at 527 g to values between @xmath104{g}$ ] and @xmath105{g}$ ] over 60ms .
this corresponds to scattering lengths between @xmath106{a_0}$ ] and @xmath107{a_0}$ ] and interaction energies between @xmath108hz and @xmath109hz , where @xmath110 is the bohr radius ( fig.3 ) .
similar to the experimental sequence presented here , one can also start from an interacting system with large tilt and both particles in the deeper well and then adiabatically go to the ground state of the symmetric double well @xcite . to confirm the adiabaticity of the ramps in tilt and in interaction strength
, we drive the system from state @xmath99 to the final configuration , hold the system for different durations and then reverse both ramps . to check for excitations ,
we measure the mean atom number in the ground state both with and without the ramps . within the statistical uncertainty of our measurement
we could not detect any heating caused by the ramps . from these measurements ,
we determine a fidelity of @xmath111% of preparing the system in the ground state of the symmetric double well .
to observe the transition from the metallic to the insulating regime , we measure the probability of single and double occupancy in the symmetric double well as a function of interaction strength ( fig.3 of main text ) .
hence , we first prepare the system either in state @xmath98 or state @xmath44 at different interaction strengths ( see sec.[sec : pass_eigenstates ] ) . then
, we quickly ramp up the potential barrier to decouple the wells and measure the atom number in either the left or the right well .
we repeat this measurement 300 times for each well and count how often we detect zero atoms ( @xmath112 ) , one atom @xmath113 ) , or two atoms ( @xmath114 ) .
the probabilities of measuring @xmath115 , @xmath116 , or @xmath80 atoms follow as : @xmath117 , where @xmath118 and @xmath119 is the total number of measurements in both wells . in our measurement
, we have finite fidelities both for the preparation of the initial state and the detection of the final atom number .
both the preparation fidelity and the detection fidelity reduce the measured atom numbers as compared to an ideal measurement . in a symmetric
double well the influence of these processes on the measured probabilities can not be distinguished .
we therefore combine the preparation and detection fidelity into an overall fidelity @xmath120 which we define as the mean atom number per well .
we determine @xmath120 by averaging over all individual measurements . due to the finite fidelity @xmath121 of our experiment , the measured probabilities @xmath122 , @xmath123 , and @xmath124 deviate from the probabilities @xmath110 , @xmath125 , and @xmath126 of an ideal experiment with @xmath127 : @xmath128 these three equations can be inverted to calculate the ideal probabilities : @xmath129 using these corrected probabilities , we define single occupancy as the probability of measuring a single atom ( @xmath130 ) and double occupancy as the sum of the probabilities of measuring either zero or two atoms ( @xmath131 ) . for comparison , the results for single occupancy and double occupancy in the ground state @xmath98 using both the corrected and the non - corrected probabilities are shown in fig.5 . to verify that our measurement of the occupation statistics is insensitive to small drifts of the tilt
, we performed the measurement for state @xmath44 at different tilts in a range of @xmath132{hz}$ ] around @xmath33 .
we observe that while the tilt affects the probabilities @xmath110 and @xmath126 of finding zero or two particles in the individual wells there is no significant change in the double occupancy @xmath133 .
to directly observe the effect of the tunnel coupling on the energy of our double - well system , we measure the energy difference @xmath134 between state @xmath44 and state @xmath45 by trap modulation spectroscopy . we start by initializing state @xmath44 with a tunnel coupling of @xmath135{hz}$ ] and interaction energies between @xmath136 and @xmath137 , as described above .
then we perform a sinusoidal modulation of the total trapping power with frequencies between @xmath138{hz}$ ] and @xmath139{hz}$ ] for a duration of @xmath140{ms}$ ] .
the atoms are transferred resonantly to state @xmath45 if the modulation frequency matches the energy difference @xmath25 . to minimize the distortion of the energy spectrum
, we modulate the tunnel coupling only with a small amplitude of @xmath141 . to determine whether the atoms have been transferred to state @xmath45
, we adiabatically turn off the interactions , ramp to a large tilt @xmath142 and measure the number of atoms in the right well .
if the system has remained in state @xmath44 during the modulation , both atoms are then in the left well . however ,
if the system has been transferred to state @xmath45 , there is a large probability to observe a single atom in the right well for @xmath142 ( fig.4 ) .
we plot the mean atom number in the right well as a function of the modulation frequency and extract the resonance frequency with a gaussian fit ( fig.6a ) .
to measure the effects of second - order tunneling , we have to control the tilt better than @xmath143 , which is approximately @xmath89{hz}$ ] at @xmath144 .
this is comparable to the residual drifts in our calibration of the tilt @xmath27 over the course of several days ( see sec.[sec : tilt ] ) .
we solve this problem by measuring the resonance frequency for different tilts around @xmath102 for each of the interaction energies .
we determine the shift @xmath145 between our set value @xmath146 and the tilt @xmath27 in the experiment for each data set by fitting a rabi function @xmath147 with @xmath148 and @xmath10 as free parameters .
we then center our data by subtracting the respective @xmath148 . for a non - interacting system ,
the energy difference @xmath25 is given by the transition frequency @xmath149 .
for an interacting system , we can still use this fit to center our data , but it does not yield the correct value for @xmath25 .
hence , we compare our measurements to a numeric solution of the hubbard model with fixed parameters @xmath11 and @xmath10 as calibrated from the tunneling measurements ( dashed curves in fig.6b ) .
we observe resonance frequencies which are systematically higher than predicted .
therefore , we fit the data with a numeric hubbard model with @xmath150 as a free parameter ( solid lines ) . we obtain @xmath151 for the balanced double - well system at different values of @xmath11 from the fitted curves ( fig.6c and fig.4 of the main text ) .
one possible cause for the deviation of the fixed - parameter hubbard model from our data is that the modulation of the trapping potential results in a modified tunnel coupling .
therefore , we scale @xmath25 and @xmath11 with the modified tunnel coupling @xmath152{hz}$ ] which we obtain from the fit at @xmath23 . | we have prepared two ultracold fermionic atoms in an isolated double - well potential and obtained full control over the quantum state of this system . in particular
, we can independently control the interaction strength between the particles , their tunneling rate between the wells and the tilt of the potential . by introducing repulsive ( attractive ) interparticle interactions
we have realized the two - particle analog of a mott - insulating ( charge - density - wave ) state .
we have also spectroscopically observed how second - order tunneling affects the energy of the system .
this work realizes the first step of a bottom - up approach to deterministically create a single - site addressable realization of a ground - state fermi - hubbard system . in the presence of strong correlations , the understanding of quantum many - body systems
can be exceedingly difficult .
one way to simplify the description of such systems is to use a discrete model where the motion of the particles is restricted to hopping between the sites of a lattice .
the paradigmatic example for this approach is the hubbard model , which reduces the physics of a quantum many - body system to tunneling of particles between adjacent sites and interactions between particles occupying the same site .
while this model captures essential properties of electrons in a crystalline solid and provides a microscopic explanation for the existence of mott - insulating and antiferromagnetic phases , many questions about this hamiltonian such as whether it can explain d - wave superfluidity are still unanswered @xcite .
a promising approach to answer these questions is to use ultracold atoms trapped in periodic potentials as quantum simulators of the hubbard model @xcite .
such experiments have been performed both in large- and small - scale systems .
degenerate gases loaded into optical lattices have been used to observe the transition to the bosonic @xcite and fermionic mott insulator @xcite .
the first observation of second - order tunneling was achieved in a small - scale system by studying the tunneling dynamics of bosonic atoms in an array of separated double wells @xcite . in a recent experiment ,
these two regimes have been connected by splitting a fermionic mott insulator into individual double wells . in this way
, the strength of the antiferromagnetic correlations in the many - body system could be determined by measuring the fraction of double wells with two atoms in the spin - singlet configuration @xcite . but
despite the observation of antiferromagnetic correlations@xcite current experiments using fermionic atoms have so far failed to reach temperatures below the critical temperature of spin ordering @xcite .
recently , new experimental techniques have been developed which allow for the deterministic preparation of few - particle systems in the ground state of a single potential well @xcite .
this makes it feasible to use ultracold atoms to study many - body physics in a bottom - up approach , i.e. to start from the fundamental building block of the system and watch how many - body effects emerge as one gradually increases the system s size @xcite . here
we report on the realization of the fundamental building block of the fermi - hubbard model at half filling , which consists of one @xmath0 and one @xmath1 particle in a spin - singlet configuration in a double - well potential . at an interaction energy of @xmath2 .
the data points show the time - averaged probability of finding a single particle ( green circles ) or a pair of particles ( blue triangles ) in the right well after initializing the system with two atoms in the left well and switching on the tunneling .
pair tunneling is resonant in a symmetric double well , while conditional single - particle tunneling occurs for a tilt of @xmath3 .
the error bars denote the @xmath4 statistical uncertainty.,scaledwidth=38.0% ] by starting from this two - site realization of the hubbard model , we can test our building block in a regime where the model can still be easily solved . in the hubbard regime , the spatial wave function @xmath5 of this two - particle system can be written in the basis @xmath6 .
these basis states are all possible combinations of the localized single - particle states @xmath7 and @xmath8 of one particle in the ground state of either the left or the right well . in this basis ,
the spatial part of the hamiltonian is @xmath9 with the tunneling matrix element @xmath10 , the on - site interaction energy @xmath11 and the energy tilt @xmath12 between the wells .
diagonalizing this hamiltonian leads to three eigenstates ( @xmath13 , @xmath14 and @xmath15 ) which are symmetric and one eigenstate ( @xmath16 ) which is antisymmetric with respect to particle exchange ( fig1b ) . to prepare our double - well system , we start with two @xmath17li atoms in different hyperfine states in the motional ground state of a single optical microtrap @xcite ( sec .
i of the supplemental material @xcite ) .
we then slowly ramp on a second potential well and thereby deform our single trap into a double - well potential . during this process
, we keep the coupling between the wells negligible and thus initialize the system in state ( @xmath18 ) where both atoms reside in the ground state of the left well .
this state is the starting point for all our measurements .
we can prepare it with a fidelity of more than @xmath19{\%}$ ] .
the predominant error is that there is only one atom in the trapping potential , while the probability to start with three atoms is @xmath20{\%}$ ] .
( shaded region ) between state @xmath15 ( orange ) and state @xmath14 ( green ) using trap modulation spectroscopy . to compare our results to the eigenenergies of the hubbard model ( fig .
1b ) , we plot the sum of @xmath21 and the energy of state @xmath14 ( @xmath22 ) .
the error bars for the @xmath4 statistical uncertainties ( sec . viii of @xcite ) are smaller than the symbols of the data points . for @xmath23 , we observe that @xmath24 , which is consistent with single - particle tunneling . for increasing repulsion ,
first - order tunneling is suppressed and @xmath25 converges to the superexchange energy @xmath26 ( see dashed line).,scaledwidth=38.0% ] since the spatial wave function of the initial state is symmetric with respect to particle exchange , the spin wave function of the two fermions is in a singlet configuration . in the experiments presented here , we do not couple the position and the spin of the particles and therefore restrict the spatial wave function of the system to the symmetric eigenstates @xmath13 , @xmath14 and @xmath15 . in a first set of experiments
, we study the tunneling dynamics of the two particles in our double - well potential to characterize the hubbard parameters @xmath10 , @xmath11 , and @xmath27 of our system . to do this , we initialize the system in state @xmath18 and abruptly reduce the height of the potential barrier , which allows the atoms to tunnel between the wells . to observe the resulting dynamics
, we let the system evolve for different durations and then freeze the spatial distribution of the atoms by quickly increasing the barrier height .
we then count the number of atoms in one of the wells by recapturing them into a magneto - optical trap and measuring their fluorescence @xcite ( sec . iii of @xcite ) . for a noninteracting system and a small tilt of the double well ( @xmath28 ) , we observe long - lived tunneling oscillations whose frequencies we can set by tuning the barrier height ( sec .
v.a , fig . s1 and fig .
s2(a ) of @xcite ) . as an example
, a typical oscillation for @xmath29 with a frequency of @xmath30{hz}$ ] is shown in figure1c . to calibrate @xmath27
, we measure the oscillation frequency for different tilts and find good agreement with the effective coupling strength @xmath31 of a two - level system ( sec . v.b and fig .
s2(b ) of @xcite ) . for an interacting system
, the interaction energy creates an effective detuning for the tunneling of a single particle . in the limit of strong interactions and a symmetric double well ( @xmath32 and @xmath33 ) , the atoms can therefore only tunnel as pairs @xcite .
however , we can restore single - particle tunneling by setting a tilt that compensates for the interaction shift @xcite .
this allows us to calibrate the on - site interaction energy @xmath11 by measuring the strength of single - particle tunneling as a function of the tilt @xmath27 for different interaction strengths ( fig.2 ) .
we find good agreement with a calculation of the interaction energy of two particles in the single well @xcite ( sec .
v.c and fig .
s3 of @xcite ) . using the parameters determined from these measurements , the two - site hubbard hamiltonian ( eq .
1 ) fully describes the oscillations of two interacting particles in our double well ( solid line in fig.1d ) . to use our system as a fundamental building block of the hubbard model , we must be able to prepare the particles in the ground state of the symmetric double well . to achieve this ,
we first initialize the system with both atoms in the left well and a tilt @xmath34 . then
, we adiabatically change the tilt to bring the system into the ground state of the symmetric double well @xcite ( sec .
vi and fig .
s4 of @xcite).{\%}$ ] . ] to show that our fundamental building block already contains the physics that is responsible for the formation of ordered phases in a many - body system , we first measure the influence of the interaction energy @xmath11 on the distribution of the two particles between the wells .
we therefore determine the probabilities @xmath35 of finding the two particles on different sites ( single occupancy ) and @xmath36 of finding both particles on the same site ( double occupancy ) by measuring the probabilities of having 0 , 1 or 2 atoms in one of the wells ( sec . vii of @xcite ) . in a noninteracting system ( @xmath23 )
, the spatial wave function of two particles in the ground state is an equal superposition of the basis states @xmath18 , @xmath37 , @xmath38 , and @xmath39 .
this leads to equal probabilities @xmath40 and @xmath41 which we observe in our measurements ( fig.3a ) . in a system with strong repulsive on - site interactions , it is energetically unfavorable to have two atoms occupying the same site .
the ground state of our system then becomes a two - particle analog to a mott - insulating state which we observe as a reduction of double occupancy .
for attractive interactions we observe an increase in double occupancy that marks the onset of a paired state .
we interpret this state as the two - particle limit of the cold gas analog of a charge - density - wave state as described in @xcite .
we perform these measurements for two different barrier heights . for the larger barrier height ( @xmath42{hz}$ ] )
, we find good agreement with the prediction of the hubbard model ( fig.3a ) .
for the smaller barrier height ( @xmath43{hz}$ ] ) , we observe a small deviation from the theoretical expectations which might indicate that our system is approaching the limits of the hubbard approximation .
since in our isolated , two - particle system excited states are stable against relaxation , we can also prepare the system in the highest - energy eigenstate @xmath44 ( fig .
s4 of @xcite ) . when measuring the occupation statistics of this state as a function of interaction strength , we find that the number of doubly occupied sites is enhanced for repulsive interactions ( fig.3b ) .
this allows us to study the charge - density wave regime in a system with repulsive interactions . in our final measurement
, we directly probe the effect of the tunnel coupling on the energy of the system by trap modulation spectroscopy ( sec . viii of @xcite ) .
for this , we initialize the system in state @xmath44 and measure the energy difference to state @xmath45 as a function of interaction strength . for @xmath23
, the atoms are delocalized over the wells by single - particle tunneling .
this leads to a change in the kinetic energy of the system that is proportional to the tunnel coupling @xmath10 ( fig.4 ) .
as we increase the interactions , the system enters the insulating regime and first - order tunneling is suppressed . however , second - order tunneling is still possible and we observe a crossover to a new energy scale given by the superexchange energy @xmath26 .
this energy is directly responsible for the appearance of spin order in the ground state of the hubbard model . by combining a series of isolated double wells we can realize a dimerized lattice @xcite where each dimer contains two fermions in a spin - singlet configuration with high fidelity . by adiabatically lowering the barriers between the double wells
, we can then prepare a low - entropy state in a homogeneous lattice .
since our systems can be prepared with arbitrary filling factors they are also ideally suited to study the effects of hole doping .
additionally , the tunability of our potential allows us to explore finite - size lattices with arbitrary geometries @xcite and introduce controlled disorder into our system .
finally , our experiments provide a starting point for scalable quantum computation with neutral atoms @xcite .
+ 33ifxundefined [ 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 10.1126/science.235.4793.1196 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.81.3108 [ * * , ( ) ] link:\doibase 10.1038/nature07244 [ * * , ( ) ] link:\doibase 10.1126/science.1165449 [ * * , ( ) ] link:\doibase 10.1126/science.1192368 [ * * , ( ) ] link:\doibase 10.1038/nature09378 [ * * , ( ) ] link:\doibase 10.1146/annurev - conmatphys-070909 - 104059 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.170403 [ * * , ( ) ] link:\doibase 10.1038/415039a [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.130403 [ * * , ( ) ] link:\doibase 10.1038/nature06112 [ * * , ( ) ] link:\doibase 10.1126/science.1150841 [ * * , ( ) ] link:\doibase 10.1126/science.1236362 [ * * , ( ) ] @noop link:\doibase 10.1088/0034 - 4885/74/5/054401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.180401 [ * * , ( ) ] link:\doibase 10.1126/science.1201351 [ * * , ( ) ] link:\doibase 10.1103/physrevx.2.041014 [ * * , ( ) ] link:\doibase 10.1126/science.1250057 [ * * , ( ) ] link:\doibase 10.1126/science.1240516 [ * * , ( ) ] @noop http://www.ub.uni-heidelberg.de/archiv/12961[ph.d .
thesis ] , ( ) link:\doibase 10.1103/physrevlett.81.938 [ * * , ( ) ] link:\doibase 10.1103/physreva.74.022712 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.075303 [ * * , ( ) ] link:\doibase 10.1126/science.1116955 [ * * , ( ) ] link:\doibase 10.1038/nature04918 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.93.186802 [ * * , ( ) ] link:\doibase 10.1103/physreva.79.033620 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/13/4/043007 [ * * , ( ) ] link:\doibase 10.1103/physreva.80.030302 [ * * , ( ) ] link:\doibase 10.1038/nature06011 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.010503 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.010502 [ * * , ( ) ] |
on hadron and strange nuclear physics , nuclear system with anti - kaons ( @xmath11 mesons = ( @xmath12 , @xmath13 ) ) has been a hot issue , since the anti - kaon is expected to cause several interesting phenomena in finite nuclear system due to a strong attraction between anti - kaon and nucleon . in particular
, the anti - kaon could be a key to access dense nuclear matter , for which partial restoration of chiral symmetry @xcite and kaon condensation @xcite have been discussed for a long time .
both theoretical and experimental studies indicate that the @xmath6 interaction in the isospin @xmath10 channel is strongly attractive . while , it is known that mass of the excited hyperon @xmath14 can not be reproduced in a naive quark model with @xmath15-wave excitation @xcite , namely , mass of @xmath14 is predicted about 100 mev larger than the pdg value exceptionally . consequently , @xmath14 is considered to be a quasi - bound state of anti - kaon and nucleon since it exists at only @xmath16 mev below @xmath6 threshold .
for example , chiral unitary model @xcite , based on a meson - baryon dynamics , has successfully explained various properties of the @xmath14 .
thus , the @xmath14 is getting recognized as a @xmath6 quasi - bound state , rather than a genuine three - quark state .
such the attractive nature of the @xmath6 interaction with a quasi - bound state , is consistent with the repulsive nature of the low - energy @xmath6 scattering data @xcite and the @xmath17-level energy shift of kaonic hydrogen atom @xcite which has been updated precisely @xcite . for the @xmath6 interaction ,
there are two kinds of potential : one is phenomenological energy - independent potentials ( _ e.g. _ ref .
@xcite ) and the other is chiral su(3)-based energy - dependent potentials @xcite . both are fitted to observables of the @xmath6 system available at this moment , and are applied and discussed intensively .
the former type of potentials is more strongly attractive than the latter type in @xmath6 sub - threshold region .
one study based on a phenomenological @xmath6 potential @xcite argues a possibility of so - called deeply bound kaonic nuclei where an anti - kaon is deeply bound in finite nuclei with a binding energy of more than 100 mev , and such a deeply - bound state could exist as a quasi - stable state since the main decay channel @xmath18 is closed . in such a state , nucleons are drawn to the anti - kaon by the strong @xmath6 attraction , and hence a dense system is generated . in case of light @xmath19-shell nuclei ,
studies with antisymmetrized molecular dynamics ( amd ) method have shown that the average density amounts to @xmath20 , when an anti - kaon is added @xcite .
( @xmath21 : normal nuclear density , around 0.17 @xmath22 . ) in case of medium and heavy nuclei , studies with the relativistic mean field ( rmf ) approach have been carried out , and it is shown that @xmath12 mesons can create a dense state inside the nucleus @xcite . therefore , we can expect kaonic nuclei to be a doorway to dense nuclear matter .
thus , kaonic nuclei are considered to be an exotic system involving several interesting aspects from the viewpoint of hadron and nuclear physics . to reveal the nature of kaonic nuclei ,
a prototype system of kaonic nuclei `` @xmath0 '' has been studied extensively.-@xmath23-@xmath24 coupled - channel system with the quantum numbers of @xmath25 and @xmath26 has been considered .
such a three - body system is denoted symbolically and representatively , as `` @xmath0 '' in this article . ] since @xmath0 is a three - body system composed of a single @xmath12 meson and two protons , various approaches are adopted . as summarized in ref .
@xcite , the resulting binding energy and decay width of @xmath0 become different depending on the combination of methods ( variational approach or faddeev - ags ) and potentials ( phenomenological or chiral - theory based ) @xcite . from that time
, further studies on @xmath0 have been carried out . a variational calculation with the hyperspherical harmonics basis function
is reported in ref .
@xcite , where the result agrees with that of an earlier study of another variational calculation @xcite when they use the same @xmath6 potential .
a faddeev - ags calculation with an energy - dependent type of chiral su(3)-based @xmath6 potential is reported in ref .
@xcite , where a @xmath0 state with small binding energy and large decay width is obtained similarly to the the variational calculation @xcite . on the other hand , there are several experiments to search for @xmath0 states . actually , the experimental result reported by finuda collaboration @xcite triggered studies of @xmath0 , although several questions were casted to their interpretation of the result @xcite .
disto collaboration @xcite reported a bump structure found in a @xmath27 invariant - mass distribution from analysis of the past data on a @xmath28 reaction .
these two collaborations claim that _ if the observed state is a @xmath0 bound state _
the @xmath0 is strongly bound with the binding energy of more than 100 mev , although its decay width is rather different between two .
thus , although much effort have been devoted to the study of @xmath0 , the definite conclusion has not been achieved yet in both the theoretical and experimental studies . however , in the theoretical side , we have one consensus that the @xmath0 exists as a resonance between the @xmath1 and @xmath23 thresholds , as commonly reported in all those calculations . from those theoretical studies , we believe that the following two ingredients are important in theoretical studies of the @xmath0 system : _ 1 .
coupled - channel problem _ and _ 2 .
resonant state_. we employ a coupled - channel complex scaling method ( cccsm ) to study the @xmath0 , since both the ingredients can be dealt with in this method simultaneously . here
, the complex scaling method ( csm ) is an established powerful tool to investigate resonant states , which has already succeeded greatly in the studies of resonant states of stable / unstable nuclei @xcite .
the csm has several advantages to investigate finite nuclear systems as follows : first , we can handle resonant states in the same way as bound states , since the resonant wave function in the csm can be represented by using only the @xmath29 basis functions such as the gaussian basis functions which have been often used in bound - state studies .
second , it is straightforward to increase the number of particles in the csm , which means that we can apply the csm to various many - body systems .
in addition , detailed properties of resonant states can be investigated by analyzing the obtained csm wave function as usually done for bound states . as the first attempt
, we have applied the cccsm to the two - body @xmath6-@xmath7 system in our previous paper @xcite .
( @xmath3 means @xmath4 and @xmath5 hyperons . ) through the study of scattering states as well as the resonant state @xmath14 , we have confirmed that the cccsm is quite useful to look into such a hadronic system . in that work ,
we have constructed a @xmath6-@xmath7 potential based on the chiral su(3 ) theory , which has a gaussian form factor in the coordinate space and the energy - dependence .
it is shown in studies based on chiral models @xcite , that @xmath14 should possess so - called a double - pole structure .
we reconfirmed such a structure with our gaussian - type potential .
we have successfully identified the lower pole as a broad resonant state , in addition to the higher pole , by using an improved gaussian basis function in the cccsm @xcite .
since we have confirmed that the cccsm is quite effective to the two - body system of @xmath6-@xmath7 , we tackle the three - body kaonic nucleus @xmath0 in this article . to study the @xmath0 , normally we have to solve an equation in the coupled - channels @xmath1 , @xmath23 and @xmath24 .
but , in this paper we propose a convenient method to reduce such a coupled - channel problem to a single - channel problem , namely , we combine the cccsm and the feshbach projection method @xcite . with this method , we can handle the @xmath1-@xmath23-@xmath24 complicated system effectively as a simple @xmath1 system without loosing effect of the decay to two other open channels .
actually , we study the @xmath0 as a gamow state and obtain the eigenstate as a definite pole on the complex - energy plane .
thus , the @xmath0 is treated as a resonant state correctly in this study .
in contrast , in the earlier studies of @xmath0 with variational approaches @xcite , the @xmath0 has been investigated within a bound - state approximation and the decay width is perturbatively estimated with the obtained wave function .
this article is organized as follows . in the next section ,
we explain our new method of cccsm+feshbach projection in detail and give all tools for the present calculation of the @xmath0 . in the section 3
, we examine our method by solving the two - body @xmath6-@xmath7 system .
main results of this paper , _
i.e. _ application of our method to the three - body @xmath0 , are shown in the section 4 .
section 5 is devoted to summary of the present study and discussion of our future plans .
here , we give a brief explanation on the usual complex scaling method ( csm ) on which the present study is based @xcite . in the csm , hamiltonian
@xmath30 and wave function @xmath31 are transformed with the complex scaling ( complex rotation ) operator @xmath32 as @xmath33 and @xmath34 , respectively . with the complex scaling the coordinate @xmath35 and the conjugate momentum @xmath36 in the hamiltonian and wave function
are transformed as @xmath37 where the variable @xmath38 is called as the scaling angle . in eigenvalues of the complex - scaled hamiltonian ,
those of scattering continuum states appear along so - called @xmath39 line on the complex - energy plane , which satisfies a relation @xmath40 .
( the variable @xmath41 means a complex eigen energy with @xmath42 . )
namely , they are dependent on the scaling angle . on the other hand , eigenvalues of bound and resonant states are proven to be independent of the scaling angle .
in addition , as is easily checked , if we choose appropriate values of the scaling angle @xmath38 , wave functions of resonant states are transformed to become square - integrable , which are originally not so .
therefore , the resonant - state wave function , which is complex - scaled , can be expanded with a square - integrable @xmath29 basis function such as gaussian basis functions , similarly to the bound - state wave functions . due to those nature of the csm
, we can obtain the eigen energies and eigen wave functions of resonant states , by diagonalizing the complex - scaled hamiltonian with gaussian basis functions .
detailed explanation on the complex scaling method is summarized in ref .
@xcite . in the present study
, we reduce a coupled - channel problem to a single - channel problem for an economical calculation , based on feshbach projection method @xcite . in the feshbach method , a model space ( @xmath15 space ) and outer space of the model space ( @xmath43 space ) are assigned with @xmath44 and @xmath45 .
schrdinger equation is given as a coupled equation of wave functions for @xmath15 and @xmath43 spaces as @xmath46 which @xmath47 and @xmath48 denote @xmath15- and @xmath43-space wave functions , respectively . by the elimination of the @xmath43-space wave function , an equation for the @xmath15-space wave function
is derived from eq .
( [ fesh_cc ] ) : @xmath49 since hamiltonian for the @xmath15 space @xmath50 is composed of the kinetic energy term @xmath51 and the potential term @xmath52 , the above equation can be written as @xmath53 here , the term @xmath54 is regarded as an effective potential for the @xmath15 space .
thus , we obtain a single - channel schrdinger equation for the @xmath15-space wave function in a formal way . in application of feshbach method to actual studies ,
the problem is how to represent the @xmath43-space green function , @xmath55 in eq .
( [ fesh_s1 ] ) .
we realize the feshbach method with help of a nature of the complex scaling method ( csm ) as follows . it is proven that the closure relation holds in the csm which includes explicitly resonant states as well as continuum scattering states and bound states @xcite .
( _ extended closure relation _ , ecr , proposed by berggren @xcite ) the ecr is shown to be useful to represent the green function of a system @xcite .
in addition , the ecr is well described approximately with a set of finite number of the complex - scaled eigenstates @xmath56 which are obtained by the diagonalization of a complex - scaled hamiltonian @xmath57 with gaussian basis functions @xmath58 @xcite : @xmath59 where @xmath60 is the state index and complex parameters @xmath61 are determined by a diagonalization of @xmath57 .
we incorporate the ecr on the @xmath43 space into the feshbach method .
first , we consider the complex - scaled green function for the @xmath43-space , @xmath62 . with the application of the ecr shown in eq .
( [ ecr ] ) , it is given approximately as @xmath63 where eigenenergies @xmath64 and eigenstates @xmath65 of the complex - scaled hamiltonian @xmath66 are calculated with gaussian basis functions @xmath58 . by the inverse transformation @xmath67
, we obtain the non - scaled green function @xmath55 from the complex - scaled one @xmath68 ; @xmath69 . substituting the obtained @xmath70 to eq .
( [ fesh_s2 ] ) , we can represent the effective @xmath15-space potential as @xmath71 where @xmath72 .
since the eigenstates @xmath65 are expanded with gaussian basis function , the effective potential is expressed with gaussian functions when the original coupled - channel potential is given in a gaussian form .
therefore , the effective potential @xmath54 derived in this way is easily handled in conventional many - body calculations with gaussian basis functions . thus , we reduce a coupled - channel problem to a single - channel problem with feshbach projection method which is assisted with a unique nature of the complex scaling method .
we call this method as a coupled - channel complex scaling method with feshbach projection , which is hereafter denoted shortly as _
`` cccsm+feshbach method''_. in theoretical studies , the @xmath0 system is treated as a coupled - channel system of @xmath1 , @xmath23 and @xmath24 , involving quantum numbers @xmath25 and @xmath26 .
we apply the cccsm+feshbach projection to the @xmath1-@xmath2 coupled - channel problem to reduce a @xmath1 single - channel problem .
( @xmath73 , @xmath5 ) at first , we consider the two - body system of @xmath6-@xmath7 .
when we set the @xmath6 channel to @xmath15 space and the @xmath7 channels to @xmath43 space , we can derive an effective @xmath6 potential for each isospin state ( @xmath74 ) with the cccsm+feshbach method : @xmath75 where the index @xmath76 indicates @xmath5 for @xmath10 and ( @xmath4 , @xmath5 ) for @xmath77 .
as explained in the previous section , eigenstates and eigenenergies @xmath78 are calculated with diagonalization of the complex - scaled hamiltonian @xmath79 for each isospin @xmath80 channel , which is @xmath81 where the channel indices @xmath82 and @xmath83 are @xmath18 for @xmath10 , and ( @xmath84 , @xmath18 ) for @xmath77 .
the terms of @xmath85 and @xmath86 are a mass of the @xmath82 channel measured from the @xmath6 threshold and the relative kinetic energy term of the @xmath82 channel which is complex - scaled , respectively .
the last term @xmath87 is a complex - scaled potential coupling between channels @xmath82 and @xmath83 with isospin @xmath80 .
note that hereafter in this article the variable `` @xmath88 '' means the scaling angle which is used to construct an effective potential by the elimination of @xmath43-space components as explained in the previous section .
with the effective @xmath6 potential @xmath89 plugged in , a three - body hamiltonian for the single @xmath1 channel is constructed to be @xmath90 the first two terms are kinetic - energy operators with respect to a jacobi coordinate , @xmath91 and @xmath92 .
the term @xmath93 is a nucleon - nucleon potential .
the last term is the effective @xmath6 potential for a @xmath94 pair with isospin @xmath80 .
detailed explanation on the @xmath8 and @xmath6 potentials will be given at the beginning of the sections [ sec : result-2body ] and [ sec : result-3body ] .
a trial wave function of the @xmath1 system with quantum numbers @xmath95 is constructed in the similar way to an earlier study with a variational approach @xcite . since the spin of the two nucleons are assumed to be zero , the trial wave function consists of two components that satisfy the antisymmetrization for two nucleons . in one component
, @xmath8 state has even - parity and isospin 1 , and in the other component it has odd - parity and isospin 0 ; @xmath96_0]_{(t , t_z)=(1/2 , 1/2)}\rangle .
\end{split}\ ] ] in the present study , the spatial part of the wave function @xmath97 is expanded with correlated gaussian basis functions @xcite , so that the @xmath8 parity is realized correctly in each part ; @xmath98 the variables @xmath99 is complex - valued parameters which are determined by the diagonalization of the complex - scaled hamiltonian . here
, the correlated gaussian function is @xmath100 , \end{split}\ ] ] where @xmath101 indicates a jacobi coordinate @xmath102 , @xmath103 is a real - symmetric @xmath104 matrix and @xmath105 is a normalization factor .
we comment that the basis functions , @xmath106 and @xmath107 , are even- and odd - parity functions for the exchange of two nucleons , respectively .
resonant states of the @xmath1 system are obtained in usual way with the complex scaling method .
the hamiltonian for the @xmath1 system , @xmath108 given in eq .
( [ h_knn ] ) , is complex - scaled with a scaling angle @xmath109 .
the complex - scaled hamiltonian , @xmath110 , is diagonalized with the basis functions @xmath111 involving spin - isospin wave functions .
it is remarked that the scaling angle used to find resonant states of @xmath1 system is denoted as `` @xmath109 '' hereafter , to distinguish from the scaling angle @xmath88 which is used for the construction of the effective potential .
the effective @xmath6 potential @xmath112 which is constructed with the cccsm+feshbach has an energy dependence . as shown in eq .
( [ ueff_kn ] ) , the potential depends on a @xmath6 energy ( @xmath113 ) which means the energy of a @xmath6 system included in a total system that we are considering .
in other words , to determine the potential strength we need to know the energy of a @xmath6 two - body system in the @xmath1 three - body system . however , such an energy of a subsystem in a total system can not be determined uniquely in principle .
we deal with the energy dependence of the effective potential , following a procedure proposed in a former study @xcite in which the same issue was considered .
we calculate a so - called kaon s binding energy @xmath114 as an auxiliary quantity , which is obtained by subtracting the @xmath8 energy from the @xmath1 energy : @xmath115 where the term @xmath116 is a complex - scaled hamiltonian for the @xmath8 system .
the @xmath8 hamiltonian is given as @xmath117 .
using the kaon s binding energy , we estimate the @xmath6 energy with two ansatz based on two extreme concepts ; @xmath118 where @xmath119 and @xmath120 are nucleon and anti - kaon masses , respectively . on one ansatz , we consider the anti - kaon as a field which carries the kaon s binding energy .
( see the left panel of fig .
[ fig_imageofpict ] . )
since the anti - kaon with the energy @xmath121 interacts with each nucleon here , the @xmath6 energy @xmath122 is equal to @xmath123 , namely @xmath124 , with static approximation applied to nucleons . on the other ansatz
, we treat the anti - kaon as a particle .
since the anti - kaon is bound by two nucleons and the kaon s binding energy is provided by them , the binding energy per a @xmath6 bond should be a half of @xmath114 .
( see the right panel of fig .
[ fig_imageofpict ] . )
therefore , the energy of a @xmath6 pair is equal to @xmath125 .
hereafter , we denote the first ansatz as _ field picture _ and the latter ansatz as _ particle picture_. for the convenience , we refer the @xmath6 energy measured from the @xmath6 threshold ; @xmath126
. means the kaon s binding energy .
details are given in the text.,title="fig:",width=240 ] means the kaon s binding energy .
details are given in the text.,title="fig:",width=240 ] when bound and resonant states of the @xmath1 system are considered with such energy - dependent potentials , the self - consistency for the @xmath6 energy has to be taken into account as explained in the former study @xcite .
the @xmath6 energy set in the effective @xmath6 potential should finally coincide with that estimated with the obtained wave function by following the above - mentioned ansatz .
it is noted that the self - consistency is realized for the _ complex _
@xmath6 energy in the current study .
we , here , treat a resonant @xmath1 state as a gamow state with the correct boundary condition .
since the pole energy on the complex - energy plane is explicitly considered , the @xmath6 energy is treated as a complex value . on the other hand ,
such a self - consistency is considered only for the real energy in the former study with a variational approach @xcite , since the @xmath1 is treated within a bound - state approximation having a real binding energy .
before the investigation of the @xmath0 system , we study the nature of our method on the two - body @xmath6-@xmath7 system . as the @xmath6(-@xmath7 ) potential @xmath127 shown in eq . ( [ ueff_kn ] ) , which is the origin of the effective @xmath6 potential , we use a chiral su(3)-based potential that was proposed in our previous study @xcite .
our @xmath6(-@xmath7 ) potential is energy - dependent and is given with a single - range gaussian form in the coordinate space . a non - relativistic version of this potential , called _ nrv2c _
, are employed in this section . [ [ test - calculation - of - cccsmfeshbach - projection - on - a - two - body - barkn - pi - y - system ] ] test calculation of cccsm+feshbach projection on a two - body @xmath6-@xmath7 system ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ .quantities of an @xmath10 @xmath6-@xmath18 system calculated with cccsm and cccsm+feshbach .
the quantities of @xmath128 and @xmath129 are @xmath6 and @xmath18 scattering lengths , respectively .
the values ( @xmath130 , @xmath131 ) indicate a resonance position of the system on the complex energy plane . @xmath132 and @xmath133 are a meson - baryon mean distance of each component of the resonant state . here
, energies and lengths are given in units of mev and fm , respectively . [ cols="<,^,^,^ " , ] we have investigated the properties of those potentials .
the pole positions of the @xmath10 @xmath6-@xmath18 resonances are given in the upper part of table [ tab_nrv2c - sm ] .
those potentials have two poles , higher pole @xmath134 and lower pole @xmath135 , in the @xmath10 channel , similarly to the nrv2c potential @xcite .
the resonant poles of the nrv2c - sm1 potential are less bound than those of the nrv2c potential . in the @xmath10 @xmath6 scattering amplitude ,
the resonance structure is certainly shifted to weaker binding side ( namely , closer to the @xmath6 threshold ) , compared with the nrv2c potential .
( see the left panel of fig .
[ fig_nrv2c - sm - i=0 ] ) in addition , the @xmath10 @xmath6 scattering length has large imaginary part in case of the nrv2c - sm1 potential .
reflecting this fact , the higher pole of the @xmath10 resonances , which is strongly coupled to the @xmath6 channel , has also large imaginary energy . by the way
, the nrv2c - sm0 potential shows the same poles and scattering amplitudes as the nrv2c potential , since the @xmath10 part of both potentials are constrained with the martin s value as explained above .
thus , it is found that our chiral su(3)-based potential becomes less attractive and more absorptive in the @xmath10 channel , when the siddharta result is taken into consideration .
scattering amplitudes calculated with nrv2c - sm1 and nrv2c potentials .
( left ) @xmath6 scattering amplitude .
( right ) @xmath18 scattering amplitude .
bold line with filled circle ( thin line ) corresponds to the scattering amplitudes of nrv2c - sm1 ( nrv2c ) potential .
black - solid ( red - dashed ) line represents the real ( imaginary ) part of the scattering amplitude .
the vertical dashed line indicates the @xmath6 threshold .
, title="fig:",width=259 ] scattering amplitudes calculated with nrv2c - sm1 and nrv2c potentials .
( left ) @xmath6 scattering amplitude .
( right ) @xmath18 scattering amplitude .
bold line with filled circle ( thin line ) corresponds to the scattering amplitudes of nrv2c - sm1 ( nrv2c ) potential .
black - solid ( red - dashed ) line represents the real ( imaginary ) part of the scattering amplitude .
the vertical dashed line indicates the @xmath6 threshold .
, title="fig:",width=259 ] scattering amplitudes calculated with nrv2c - sm1 , nrv2c - sm0 and nrv2c potentials .
( left ) @xmath6 scattering amplitude .
( middle ) @xmath18 scattering amplitude .
( right ) @xmath84 scattering amplitude .
bold line with filled circles , bold line with crosses and thin line correspond to the scattering amplitudes of nrv2c - sm1 , nrv2c - sm0 and nrv2c potentials , respectively .
, title="fig:",width=182 ] scattering amplitudes calculated with nrv2c - sm1 , nrv2c - sm0 and nrv2c potentials .
( left ) @xmath6 scattering amplitude .
( middle ) @xmath18 scattering amplitude .
( right ) @xmath84 scattering amplitude .
bold line with filled circles , bold line with crosses and thin line correspond to the scattering amplitudes of nrv2c - sm1 , nrv2c - sm0 and nrv2c potentials , respectively .
, title="fig:",width=182 ] scattering amplitudes calculated with nrv2c - sm1 , nrv2c - sm0 and nrv2c potentials .
( left ) @xmath6 scattering amplitude .
( middle ) @xmath18 scattering amplitude .
( right ) @xmath84 scattering amplitude .
bold line with filled circles , bold line with crosses and thin line correspond to the scattering amplitudes of nrv2c - sm1 , nrv2c - sm0 and nrv2c potentials , respectively .
, title="fig:",width=182 ] as for the @xmath77 channel , the nrv2c - sm0 potential has large imaginary part of the @xmath6 scattering length , compared with the nrv2c - sm1 and nrv2c potentials , although all three potentials give the similar real part of the @xmath6 scattering length .
( lower part of table [ tab_nrv2c - sm ] ) fig .
[ fig_nrv2c - sm - i=1 ] depicts the @xmath6 , @xmath18 and @xmath84 scattering amplitudes with @xmath77 .
the nrv2c - sm1 potential gives quite the same amplitudes as the nrv2c potential .
( the amplitudes of both potentials are overlapped in the figure . )
the @xmath6 amplitude of nrv2c - sm0 potential is slightly attractive than that of other potentials .
however , the difference of scattering amplitudes between three potentials is tiny . therefore , the property of our potential in the @xmath77 channel is found not to change so much with the siddharta result .
a. dot@xmath136 , h. horiuchi , y. akaishi and t. yamazaki , phys .
b * 590 * , 51 ( 2004 ) ; phys .
c * 70 * , 044313 ( 2004 ) .
d. gazda , e. friedman , a. gal and j. mares , phys .
c * 76 * , 055204 ( 2007 ) ; phys .
c * 77 * , 045206 ( 2008 ) .
a. ciepl , e. friedman , a. gal , d. gazda and j. mares , phys .
c * 84 * , 045206 ( 2011 ) .
y. ikeda and t. sato , phys .
c * 76 * , 035203 ( 2007 ) .
n. v. shevchenko , a. gal , j. mares , and j. rvai , phys .
c * 76 * , 044004 ( 2007 ) . t. yamazaki and y. akaishi , phys .
c * 76 * , 045201 ( 2007 ) . | kaonic nuclei ( nuclear system with anti - kaons ) have been an interesting subject in hadron and strange nuclear physics , because the strong attraction between anti - kaon and nucleon might bring exotic properties to that system . in this article
, we investigate @xmath0 as a prototype of kaonic nuclei . here , @xmath0 is a three - body resonant state in the @xmath1-@xmath2 coupled channels , where @xmath3 represents hyperons @xmath4 and @xmath5 . in order to treat resonant states in a coupled - channel system properly ,
we propose newly a coupled - channel complex scaling method combined with the feshbach projection , namely cccsm+feshbach method . in this method ,
the feshbach projection is realized with help of so - called the extended closure relation held in the complex scaling method , and a complicated coupled - channel problem is reduced to a simple single - channel problem which one can treat easily .
first , we confirm that the cccsm+feshbach method completely reproduces results of a full coupled - channel calculation in case of two - body @xmath6-@xmath7 system .
we then proceed to study of three - body @xmath1-@xmath2 system , and successfully find solutions of the @xmath0 resonance by imposing self - consistency for the complex @xmath6 energy .
obtained binding energy of @xmath0 is well converged around 27 mev , with an energy - dependent @xmath6(-@xmath7 ) potential based on the chiral su(3 ) theory , independently of ansatz for the self - consistency .
this binding energy is small as ones reported in earlier studies based on chiral models . on the other hand ,
decay width of @xmath0 strongly depends on the ansatz .
we calculate also the correlation density of @xmath8 and @xmath6 pairs by using the obtained complex - scaled wave function of the @xmath0 resonance .
effect of the repulsive core of @xmath8 potential is seen in the @xmath8 correlation density . in the @xmath6 correlation density
, we can confirm survival of @xmath9 resonance ( @xmath10 @xmath6 resonance ) in the three - body resonance . |
quantum plasmonics is an emerging area of research which involves the study of the optical properties of hybrid photonic structures incorporating both plasmonic nanostructures and quantum emitters @xcite , such as atoms , molecules and semiconductor quantum dots .
these complex hybrid are active photonic structures and expected to enhance optical response significantly , for example modification of the linear susceptibility @xcite and the enhancement of nonlinear susceptibilities in several quantum systems with different level structures coupled to various plasmonic nanostrcutures @xcite .
four wave mixing is one of the above mentioned nonlinear process of light - matter interactions in which three incoming waves , indicated as @xmath0 , @xmath1 , @xmath2 in the material generate a fourth wave of frequency @xmath3 @xcite . assuming that the three incident waves have frequencies in the visible or near - infrared range , the incoming electric fields @xmath4 @xmath5with @xmath6 interact with the material s electrons to induce a nonlinear polarization @xmath7 in the illuminated volume .
the magnitude of the polarization is determined by the strength of the incident fields and the efficiency with which the material can be polarized @xcite .
the latter is indicated with the third - order nonlinear susceptibility @xmath8 , a measure of the material s response to the incoming fields .
four wave mixing ( fwm ) has found numerous practical applications , including : optical processing ; nonlinear imaging ; real - time holography and phase - conjugate optics ; phase - sensitive amplification ; and entangled photon pair production @xcite .
in several recent studies , the modification of @xmath8 ( fwm process ) susceptibility in a quantum dot system coupled to spherical nanoparticle has been investigated when the hybrid structures interacts with a weak probe field and a strong pump field @xcite .
all these works have shown for different distance between the quantum dot and the metal nanoparticle the @xmath8 susceptibility can be either enhanced or strongly suppressed .
in addition , bistable behavior has been also reported in these kind of systems @xcite . + here ,
we propose a method for increasing the efficiency of fwm processes by exploiting gold grating @xcite .
narrow peaks are observed in the transmission spectra of p - polarized light passing through a thin gold film that is coated on the surface of a transparent diffraction grating .
the spectral position and intensity of these peaks can be tuned over a wide range of wavelengths by simple rotation of the grating @xcite .
the wavelengths where these transmission peaks are observed correspond to conditions where surface plasmon resonance occurs at the gold - air interface .
light diffracted by the grating couples with surface plasmons in the metal film to satisfy the resonant condition , resulting in enhanced light transmission through the film .
the paper is organized as follows . in section
ii , we describe the fwm process in the coupled system of gold grating with a quantum oscillator . in the same section
, we introduce the hamiltonian for hybrid system .
fwm process is also included in the second quantized hamiltonian .
we derive the equations of motion for the system using the density matrix formalism for the quantized quantum oscillator .
we use phenomenological way to include damping of gold grating modes as well as quantum emitter . in section ii b , we demonstrated that fwm process can be suppressed for @xmath9 . in section ii c , we present a contrary effect where fmw process can be enhanced .
this is due to cancellation of non resonant terms in denominator . in section iii
, we further investigate the case of gold grating coupled to two quantum emitters ( both quantum emitters coupled to each other also ) simultaneously .
we conclude our results in section iv .
the total hamiltonian @xmath10 for the described system can be written as sum of the energy of the quantum oscillator @xmath11 , ( in our case we have taken a qd of energy levels @xmath12and @xmath13 ) , enegy of the elctromagnetic modes of gold grating @xmath14 for a particular angle of incidence of pump lasers @xmath15 , the interaction of the quantum oscillator with the grating modes @xmath16 @xmath17 @xmath18 @xmath19 here , we have considered that level spacing of the qd is only resonant to @xmath20 mode ( i.e. @xmath21 ) . as well as the energy transferred by the pump source @xmath22 and @xmath23@xmath24 @xmath25 for the process of @xmath26as mentioned in prl 103 , 266802 . in eq .
( 1 ) , @xmath27 ( @xmath28 ) is the excited ( ground ) state energy of the quantum oscillator .
states @xmath29 and @xmath30 corresponds to excited and ground levels of the quantum oscillator respectively .
@xmath31 are the gold grating modes at a particulat angle of incidence @xmath32 .
@xmath33 is the coupling matrix element between the field of grating mode and the quantum oscillator .
( 4 ) describes the driving the electromagnetic field modes of grating @xmath34 with @xmath35 and @xmath36 respectively .
eq.(5 ) describes where the four wave mixing takes place in which @xmath37 mode contributes two photons and @xmath38 mode single photons in the process .
we use the commutation relations @xmath39\ ] ] for deriving equations of motions . after obtaining the dynamics in the quantum approach ,
we carry @xmath40to classical expectation values @xmath41we also introduce the decay rates for @xmath42 quantum oscillator is treated within the density matrix approach .
the equations of motion take the form @xmath43 @xmath44 @xmath45 @xmath46 @xmath47 where @xmath48,@xmath49are the damping rates of the electromagnetic modes of the gold grating @xmath50and @xmath51 @xmath52are the diagonal and off - diagonal elements of the quantum oscillator respectively .
the constraints of the conservation probability @xmath53 accompanies above set of equations .
besides the time - evolution simulations , one may gain the understanding by seeking solutions of the following form . for long time behavior
we take solutions of the form @xmath54 ( condition of four wave mixing process ) , @xmath55 here we have considered that level spacing of the qd is only resonant to @xmath56 mode ( i.e. @xmath21 ) , @xmath57 inserting the solutions in above set of equations ( 7a-7e ) we have for long time behavior @xmath58 \tilde{\alpha}% _ { 1}+2i\chi ^{\left ( 3\right ) } \tilde{\alpha}_{1}^{\ast } \tilde{\alpha}_{2}% \tilde{\alpha}_{3}=\epsilon _ { p } \label{8a}\ ] ] @xmath59 \tilde{\alpha}_{2}+i\chi ^{\left ( 3\right ) } \tilde{\alpha}_{3}^{\ast } \tilde{% \alpha}_{1}^{2}=\epsilon _ { p}^{\prime } \label{8b}\ ] ] @xmath60 \tilde{\alpha}_{3}+i\chi ^{\left ( 3\right ) } \tilde{\alpha}_{2}^{\ast } \tilde{\alpha}_{1}^{2}=-if\tilde{\rho}_{ge } \label{8c}\ ] ] @xmath61 \tilde{\rho}_{ge}=if\tilde{\alpha}_{3}\left ( \tilde{\rho}_{ee}-% \tilde{\rho}_{gg}\right ) \label{8d}\ ] ] @xmath62 using equations ( 8c ) and ( 8d ) , we obtain the steady state value for @xmath63 as follows @xmath64 } -\left [ i\left ( \omega _ { 3}+\omega ^{\prime } -2\omega \right ) + \gamma _ { 3}\right ] } \ ] ] where @xmath65is the steady state value of the population inversion . if the quantum oscillator is tuned around @xmath66 @xmath67 can be suppressed .
we can see from eq .
( 9 ) that @xmath68 can attain huge values on resonance @xmath69as well as linewidth of the quantum oscillator @xmath70 is very small compared to the all other frequencies . if @xmath71 , the largeness of the @xmath72 term dominates the denominator
this results in the suppression of the generation of the fwm mode @xmath63 in our model hamiltonian system . in fig.1
we have shown that fwm process can be suppressed very effectively by coupling to gold grating to quantum oscillator .
we have time evolve eqs .
@xmath73 to obtain steady state values for the fwm intensity.without the presence of quantum oscillator , the fwm would be maximum @xmath74 when the fwm mode is on resonance @xmath75 . in fig .
1 we observe that even at the presence of this resonance condition @xmath76 , eit suppresses the fwm by 10 order of magnitude . furthermore , in this case population of excited level of quantum oscillator is maximum at this point as shown in fig.2 as well as population inversion is approximately @xmath77 similar to suppression phenomena , the interference effects can be arranged in such a way that fwm process can be carried closer to resonance . in the denominator of eq.(9 ) , the imaginary part of the first term@xmath78 } $ ] can be arranged to cancel the@xmath79 factor in the second term of the denominator .
this gives the condition @xmath80
= 0\ ] ] eq.(10 ) has two roots .
@xmath81 the first smaller root @xmath82 is not very useful for fwm enhancement , as it enhance the real part of the term @xmath83 } ~ $ ] to rapidly diverge as we have seen in suppression condition for fwm , whereas @xmath84 minimizes the absolute value of the denominator of eq . ( 9 ) that gives enhancement of fwm process . for the case of suppression of fwm
, one can safely use the approximation @xmath85 @xmath86 because excitations are suppressed in the hybrid system @xmath87 and this leads to @xmath85 = @xmath88 .
however , in case of fwm enhancement , one can not approximate @xmath85 @xmath86
. nevertheless , eq.(11 ) still serves at least a guess value for the order of @xmath89 , where fwm enhancement arises .
in case of two coupled qds we have total hamiltonian as follows .
@xmath90 @xmath18 @xmath91\ ] ] here , we have considered that level spacing of both the qds is only resonant to @xmath56 mode ( i.e. @xmath92 , @xmath93 ) .
@xmath94\ ] ] as well as energy transferred by pump source with @xmath35 and @xmath95 respectively .
@xmath96 @xmath97 is the four wave mixing hamiltonian for the process of @xmath98 as mentioned in prl 103 , 266802 . by using the commutation relation eq.(6 )
as well as proceeding like the same way for a single qd case , we get following equations for the case of coupled qds .
@xmath99 @xmath100 @xmath101 @xmath102 @xmath103 @xmath104 @xmath105 where @xmath48,@xmath49are the damping rates of the electromagnetic modes of the gold grating @xmath106 , @xmath107 and @xmath108are the diagonal and off - diagonal decay rates of the first and second quantum emitter respectively .
the constraints of the conservation probability @xmath109 and @xmath110 accompanies above set of eqs.(18a-18 g ) . in our simulation for enhancement process of fwm , we time evolve eqs .
( 18a-18 g ) numerically to obtain the long time behaviors of @xmath111 and @xmath112 .
we determine the values to where they converge when the drive is on for long enough times .
we perform this simulations for different parameter sets @xmath113 with the initial condition @xmath114 @xmath115 .
besides the time - evolution simulations , one may gain the understanding by seeking the solutions of the following form : @xmath116 , @xmath117 , @xmath118 ( condition of four wave mixing process ) , @xmath119 , @xmath120 .
here we have considered that level spacing of both the qds is only resonant to @xmath56 mode ( i.e. @xmath121 , @xmath122 and @xmath123.inserting the solutions in above set of eqs .
( 18a-18 g ) we have the following closed set of equations for the steady state dynamics @xmath58 \tilde{\alpha}% _ { 1}+2i\chi ^{\left ( 3\right ) } \tilde{\alpha}_{1}^{\ast } \tilde{\alpha}_{2}% \tilde{\alpha}_{3}=\epsilon _ { p } \label{19a}\ ] ] @xmath59 \tilde{\alpha}_{2}+i\chi ^{\left ( 3\right ) } \tilde{\alpha}_{3}^{\ast } \tilde{% \alpha}_{1}^{2}=\epsilon _ { p}^{\prime } \label{19b}\ ] ] @xmath60 \tilde{\alpha}_{3}+i\chi ^{\left ( 3\right ) } \tilde{\alpha}_{2}^{\ast } \tilde{\alpha}_{1}^{2}=-if_{1}\tilde{\rho}_{ge}^{\left ( 1\right ) } --if_{2}% \tilde{\rho}_{ge}^{\left ( 2\right ) } \label{19c}\ ] ] @xmath124 \tilde{\rho}_{ge}^{\left ( 1\right ) } = if_{1}^{\ast } \tilde{\alpha}_{3}y_{1}+ig^{\ast } y_{1}\tilde{\rho}% _ { ge}^{\left ( 2\right ) } \label{19d}\ ] ] @xmath125 \tilde{\rho}_{ge}^{\left ( 2\right ) } = if_{2}^{\ast } \tilde{\alpha}_{3}y_{2}+igy_{2}\tilde{\rho}% _ { ge}^{\left ( 1\right ) } \label{19e}\ ] ] @xmath126 @xmath127 where @xmath128 and @xmath129are constants independent of time .
@xmath130are the population inversion @xmath131 for both qds . using eqs.(19d ) and ( 19e ) in eq.(19c ) , we obtain the steady state value for @xmath132 as follows . @xmath133 where the short hand notations are @xmath134 , ~\varepsilon _ { 2}=\left [ i\left ( \omega _ { 2}-\omega ^{\prime }
\right ) + \gamma _ { 2}\right ] , ~\varepsilon _ { 3}=% \left [ i\left ( \omega _ { 3}+\omega ^{\prime } -2\omega \right ) + \gamma _ { 3}% \right ] ~$ ] and @xmath135 ~ $ ] and @xmath136 .$ ] in case of a single qd coupled to the gold grating , @xmath137 and @xmath138
, we get the steady state value of @xmath139 from eq.(20)as @xmath140}-\left [ i\left ( \omega _ { 3}+\omega ^{\prime } -2\omega \right ) + \gamma _ { 3}\right]}\ ] ] which coincides exactly with eq.(9 ) where the imaginary part of the first term @xmath141 } $ ] can be arranged to cancel the @xmath142 factor in the second term of the denominator and this gives enhancement of fwm as also discussed in previous section also . as compared to single qd case , the denominator of eq.(20 ) can ( in principal )
be arranged down to very low values in order to enhance @xmath67 to much higher values . in this case
, denominator has 3 complex @xmath143 and 2 real @xmath144 parameters which can be tuned independently .
we obtain nearly @xmath145 times enhancement by comparing the steady state values of @xmath146 that is the intensity of fwm process calculated from time evolution of eqs.@xmath147 for the chosen set of parameters as shown in fig.@xmath148 . here , frequency of second qd is kept constant and first one is varying . for the decay rates of grating modes in between @xmath149 , we get enhancement in fwm intensity around 1200 - 1600 times as compared to case of single qd discussed in previous section .
it is well demonstrated that the presence of a quantum emitter with a smaller decay rate changes the optical response of coupled grating dramatically . due to the destructive interference of the ( hybridized ) absorption paths , four wave mixing(fwm ) process
can be suppressed at the resonance frequency of the quantum emitter .
we demonstrate that a similar path interference effect can be adopted to both suppress and enhance the nonlinear four wave mixing processes ( fwm ) in a grating surface .
a quantum emitter is coupled with the electromagnetic modes of a gold grating .
we found that the fwm process can be suppressed over 10 orders of magnitude .
such an suppression can be achieved by carefully choosing the coupling strengths and the energy level spacing for quantum emitters .
when @xmath150 , the fwm intensity can be suppressed by several order of magnitude with respect to resonant value . on the other hand , the similar interference effects can be also used to enhance the nonlinear fwm intensity .
the level spacing of the single quantum emitter can be arranged so that the nonresonant terms get canceled . in case of two coupled quantum emitters by arranging energy level spacing for quantum emitters in the same way like single quantum emitter , we have enhancement in fwm intensity upto the order of @xmath151 . 99 m.s.tame ; m.s .
kim , nature physics * 9 * , 329 - 340 ( 2013 ) .
z. lu ; k. d. zhu , j. phys .
b. * 42 * , 0155502 ( 2009 ) .
a. hatef ; m.r .
singh , phys .
a. * 81 * , 063816 ( 2010 ) .
sadeghi , nanotechnology .
* 21 * , 455401 ( 2010 ) .
e. paspalakis ; s. evangelou ; v. yannopapas ; a.f .
terzis , phys . rev . a * 88 * , 053832 ( 2013 ) .
tasgin , nanoscale . * 5 * , 8616 ( 2013 ) .
z. lu ; k. d. zhu , j. phys .
b. * 41 * , 185503 ( 2008 ) .
y pu ; r grange ; c.l .
hsieh ; d. psaltis ; phys .
104 * , 207402 ( 2010 ) . i. thanopulous
; e. paspalakis ; v. yannopapas ; phys . rev .
b * 85 * , 035111 , 2012
singh , nanotechnology * 24 * , 125701 ( 2013 ) .
e. paspalakis ; s. evangelou ; a.f .
terzis , j. appl . phys . * 115 * , 083106 ( 2014 ) . b. yurke ; d. stoler , phys . rev .
a * 35 * , 4846 ( 1987 ) .
y. wang ; chia - yu lin ; a. nikolaenko ; v. raghunathan and eric o. potama , advances in optics and photonics * 3 * , 1 - 52 , 2011 .
scully and m.s .
zubairy , quantum optics , cambridge university press ( 2007 ) .
b . li ; n .- c .
kim ; m .- t .
cheng ; l. zhou ; z .- h . hao ; q .- q .
wang , opt .
express * 20 * , 1856 - 1861 ( 2012 ) .
liu ; d.z.yao ; h.m .
zhou ; f. chen ; g.g .
xiong , appl .
b. * 113 * , 603 - 610 ( 2013 ) .
li ; k .- d .
zhu , crit .
solid state mater .
* 39 * , 25 - 45 ( 2014 ) .
j. renger ; r. quidant ; n. hulst ; s. palomba and l. novotny , phys . rev .
lett . * 103 * , 266802(2009 ) .
e. poutrina ; c. cirac ; d. j. gauthier ; and david r. smith , optics express * 20 * , 11005(2012 ) .
bipin k. singh and andrew c. hillier , anal .
chemistry * 180 * , 3803 - 3810(2008 ) . | we have shown that four wave mixing ( fwm ) processes of electromagnetic field modes of a grating can be controlled by the presence of interactions with a quantum dot or a molecule made by coupled quantum dots . by choosing the appropritae level spacing for the quantum emitter
, one can either suppress or enhance the four wave mixing process .
we revel theoretically the underlying mechanism for this effect .
( i ) suppression in fwm intensity occurs simply because induced electromagnetic induced transparency does not allow the excitation at converted fwm frequency .
( ii ) enhancement emerges since fwm process can be brought to resonance .
path interference effect cancels the nonresonant frequency terms .
furthermore , we have also shown that in case of coupled quantum dots enhancement increases significantly as compared to the case of a single quantum dot . |
the quasi two - dimensional molecular series @xmath1-(bedt - ttf)@xmath5 exemplifies the complex interplay of collective phenomena that occur in correlated electron systems@xcite . in particular ,
the proximity of superconducting @xmath1-(bedt - ttf)@xmath2cu[n(cn)@xmath2]br to a mott transition , along with possible pseudogap physics and molecular disorder effects , has led to this material being extensively studied in recent years@xcite .
despite this intense interest , several experimental observations in this system lack a conclusive explanation @xcite and more experimental study is warranted . in this paper
we present a muon - spin relaxation ( @xmath0sr ) investigation of @xmath1-(bedt - ttf)@xmath2cu[n(cn)@xmath2]br .
our focus is the penetration depth in the vortex state for @xmath7 and the fluctuations of the local magnetic field in the normal state .
the layered structure of the @xmath1-(bedt - ttf)@xmath5 family of organic molecular metals is shown in fig .
[ picture](a ) .
the bedt - ttf molecules dimerize , forming molecular units which stack on a triangular lattice in two - dimensional planes .
the removal of one electron from each bedt - ttf dimer causes the tight - binding band to be half filled and in order to balance the charge layers of anion , x , are located between the partially oxidized bedt - ttf sheets .
the properties of the system are strongly dependent on the transfer integral @xmath8 which is controlled by the dimer separation and may be manipulated by applying pressure or by changing the identity of the anion@xcite .
the phase diagram of the family [ fig .
[ picture](b ) ] shows that the system may be tuned from mott insulator through superconductivity into a normal metallic state as a function of @xmath9 , where @xmath10 is the on - site coulomb repulsion ( which is a property of the dimer and is almost independent of the anion or pressure ) .
importantly , this tuning is achieved without the need for chemical doping ( in contrast to the cuprates ) and therefore minimizes structural disorder effects . at small @xmath9
coulomb correlations dominate and @xmath1-(bedt - ttf)@xmath5 with @xmath11cu[n(cn)@xmath2]cl is a mott insulator @xcite . as pressure
is increased or @xmath12 is changed to @xmath11cu[n(cn)@xmath2]br there is an insulator - to - superconductor transition @xcite .
the @xmath11cu(ncs)@xmath2 compound exhibits a larger @xmath9 still and remains a superconductor with a slightly depressed @xmath4 .
superconducting @xmath1-(bedt - ttf)@xmath2cu[n(cn)@xmath2]br ( @xmath13 k ) displays several properties [ fig .
[ master ] ] whose explanation remains obscure .
these may be broadly grouped three temperature regions . in region
i ( @xmath14 k ) there are sharp changes in the temperature dependence of the lattice constants@xcite [ fig .
[ master](d ) ] and , above 70 k , a rapid increase in the thermal expansion coefficient perpendicular to the planes@xcite @xmath15 [ fig .
[ master](e ) ] .
these coincide with a rounded maximum in the resistivity@xcite @xmath16 [ fig .
[ master](b ) ] .
the properties of this region have been linked to a glass - like freezing of terminal ethylene groups on the bedt - ttf molecules@xcite around @xmath17 k , although is has been suggested @xcite that this freezing alone can not account for the observed structural changes . in region
ii ( @xmath18 k ) nmr measurements @xcite yield a maximum in @xmath19 around @xmath20 k [ fig . [ master](c ) ] .
upon cooling through this region there is a crossover in transport properties from `` bad metal '' behavior to a more conventional fermi - liquid regime@xcite and a plateau in @xmath15 is also seen [ fig . [ master](e ) ] .
these phenomena have been linked to pseudogap physics although the details remain unclear@xcite . in region
iii ( @xmath21 k ) a vortex nernst signal is observed @xcite .
it was argued@xcite that the proximity of the @xmath11cu[n(cn)@xmath2]br material to the mott state in the phase diagram in fig .
[ picture](b ) , results in vortex fluctuations persisting above @xmath4 and , furthermore , this remnant of fluctuating superconductivity may be consistent with the occurrence of phase fluctuations in the superconducting order parameter close to the mott boundary .
this paper is structured as follows : after discussing experimental details in section ii we present the results of transverse field muon - spin relaxation measurements of the superconducting penetration depth of @xmath1-(bedt - ttf)@xmath2cu[n(cn)@xmath2]br in section iii .
section iv describes the use of longitudinal field measurements to investigate spin fluctuations in the normal state of the material and our conclusions are presented in section v.
muon - spin spectroscopy@xcite is a sensitive means of probing the superconductivity and local magnetism of molecular materials of this sort@xcite .
transverse - field ( tf ) and longitudinal - field ( lf ) muon - spin relaxation ( @xmath0sr ) measurements were made on a mosaic sample of @xmath1-(bedt - ttf)@xmath2cu[n(cn)@xmath2br ] at the isis facility , rutherford appleton laboratory , uk and the swiss muon source ( s@xmath22s ) , paul scherrer institut , ch .
the polycrystalline sample was made of @xmath23 small crystallites which grow as platelets whose large faces are parallel to the conducting layers ( the @xmath24 planes for this material ) .
these were arranged on a piece of ag foil to cover an area of @xmath25 @xmath26 so that the conducting layers are parallel to the plane of the mosaic .
tf measurements were made using the musr spectrometer at isis where the sample was mounted on a hematite backing plate in a helium cryostat with the sample oriented at @xmath27 to both the applied magnetic field and the initial muon spin . in order to reduce any effect of ethylene disorder on the superconducting properties ,
the sample was slow cooled at a rate of @xmath28 k / hour .
tf measurements were also made using the gps instrument at s@xmath22s where the sample plane was oriented at @xmath29 to the applied magnetic field and no slow cooling procedure was followed .
lf measurements were made on the argus spectrometer ( isis ) in the latter @xmath29 geometry .
tf @xmath0sr provides a means of accurately measuring the internal magnetic field distribution in a material , such as that due to the vortex lattice ( vl ) in a type ii superconductor @xcite . in a tf @xmath0sr experiment spin polarized muons
are implanted in the bulk of a material in the presence of a magnetic field @xmath30 , which is applied perpendicular to the initial muon spin direction .
muons stop at random positions on the length scale of the vl where they precess about the total local magnetic field @xmath31 at the muon site with frequency @xmath32 , where @xmath33 mhz t@xmath34 .
the observed property of the experiment is the time evolution of the muon - spin polarization @xmath35 , which allows the determination of the distribution @xmath36 of local magnetic fields across the sample volume via @xmath37 where the phase @xmath38 results from the detector geometry . above @xmath4
some broadening of the spectrum is caused by randomly directed nuclear moments relaxing the muon spins .
below @xmath4 the spectrum broadens considerably due to the contribution from the vl , but this component was found to be only @xmath39% of the total signal for @xmath7 . allowing for part of the non - superconducting signal to come from the cryostat and sample mounting material , we estimate that at most 50% of the sample is giving a superconducting signal .
this is significant since it suggests that the majority of muons are not stopped in regions where there is a well defined vl and may indicate the coexistence of a competing phase in this material .
in fact , there is evidence for a phase separation between superconducting and insulating regions in this material@xcite which is strongly dependent on disorder , such as that introduced through rapid cooling . the non - superconducting contribution to our signal for @xmath7 decreases weakly and linearly with increasing temperature .
rapid cooling of the sample was found to reduce the superconducting volume fraction still further .
with such a small superconducting fraction it is difficult to accurately extract detailed lineshape information from our data , however the rms width of the vl field distribution @xmath40 can be straightforwardly obtained by fitting the data to the sum of two gaussian relaxation components , reflecting the superconducting vl and non - superconducting contributions .
these are easily separated below @xmath41 since the contribution from the vl has a much larger relaxation rate .
the field dependence of the vl component is shown in fig .
[ tffig](a ) .
the low field data ( 40 mt and below ) fit well to the width expected from the solution of a ginzburg - landau ( gl ) model of the triangular vl field distribution @xcite with an in - plane penetration depth of 0.47(1 ) @xmath22 m .
@xmath40 is not very sensitive to @xmath42 in this field range , so we assume the reported value @xcite @xmath42=10(2 ) t ; the gl parameter @xmath1 is then estimated to be 80(10 ) .
the penetration depth in the highly conducting planes can also be estimated via the approximate expression for a high anisotropy superconductor @xmath43^{\frac{1}{2 } } , \label{bg}\ ] ] where @xmath44 is the flux quantum and @xmath45 is the angle of the applied field with respect to the normal to the plane .
this field - independent expression is valid at intermediate fields @xmath46 and the @xmath40 value obtained from eq .
( [ bg ] ) for @xmath47 = 0.47 @xmath22 m is shown as the dashed line in fig .
[ tffig](a ) ; this is seen to be reasonably close to the field - dependent gl width in a field region centred around 10 mt . [ cols="^,^,^,^,^,^,^,^",options="header " , ] our value of @xmath47 is rather smaller than previous reports , e.g. the range @xmath48 m obtained from reversible high field magnetization measurements@xcite , and it is significantly lower than the value @xmath49 m estimated from a previous @xmath0sr measurement @xcite in a field of 300 mt .
these differences can be understood by noting that the measured linewidth decreases sharply for @xmath50 mt ( fig .
[ tffig](a ) ) . if this suppressed linewidth were taken to represent a full 3d vl , a penetration depth of order @xmath51 m would also be obtained from our data using eq .
( [ bg ] ) , in good agreement with the previous @xmath0sr result .
the decrease in @xmath40 with applied field that we observe is not surprising , given the vortex phase diagram in ( bedt - ttf)@xmath2cu(scn)@xmath2@xcite . in that material the ideal 3d triangular vl which exists at low temperatures and low fields
@xcite is destroyed by the application of fields above a disorder - dependent threshold , that can be as low as 6 mt , causing a transition to a decoupled - layer vortex glass phase , accompanied by a sharp decrease in the measured linewidth @xcite .
it is therefore likely that a similar transition to a decoupled - layer vortex glass phase occurs in the @xmath11cu[n(cn)@xmath2br material above a theshold field in the 4080 mt region .
the effect on the width of the field distribution in the limit of losing all inter - layer correlation has been calculated@xcite , giving @xmath52 where @xmath53 is the inter - layer spacing and @xmath54 is vortex spacing within the layers .
this limit is shown as the dot - dash line in fig .
[ tffig](a ) and it can be concluded that some residual interlayer correlations remain here since the data points lie significantly above this limit .
a further feature of our data in the decoupled vortex phase is that the superconducting fraction recovers to the larger value obtained in measuremants made under slow - cooling conditions , suggesting that the disorder - induced phase separation is becoming significantly reduced once the layers are decoupled .
the temperature dependence of @xmath40 has been measured in a field of 3 mt applied in a 45@xmath55 geometry and this is shown in fig . [ tffig](b ) .
this can be fitted to an empirical power law @xmath56 , % $ \label{one}\ ] ] where @xmath57 is the zero temperature contribution from the vl .
the vl here therefore appears to thermally stable , in contrast to the case of ( bedt - ttf)@xmath2cu(scn)@xmath2 where a clear melting of the vl is seen @xcite .
the values @xmath57=0.31(1 ) mt and @xmath41=11.3(4 ) k are obtained from the fit .
the penetration depth in the highly conducting planes could be obtained by taking @xmath58 in eq .
( [ bg ] ) , however the absolute accuracy is reduced compared to 90@xmath55 measurements for two reasons : firstly the width measured at @xmath27 is particularly sensitive to any angular misalignment between the sample and field and secondly the angular scaling in highly anisotropic superconductors can be much more complex than the simple form implied by eq .
( [ bg ] ) @xcite . returning to the 90@xmath55 data , we note that a marginally better fit is obtained by allowing for an asymmetric lineshape @xcite parameterized using the skewness parameter @xmath59 .
the optimum fit is achieved with @xmath60 , which is reasonably close to the ideal triangular lattice value of 0.60 , that was observed in more detailed studies of the vortex phases of the x = cu(ncs)@xmath2 compound @xcite .
this fit also yields 0.47(1 ) @xmath22 m for @xmath47 .
the values obtained here and in previous muon studies for 90@xmath55 geometry are summarized in table [ table1 ] .
we may compare our values of @xmath47 and @xmath4 for the x = cu[n(cn)@xmath2]br material to values obtained previously for the x = cu(scn)@xmath2 material .
this is shown as the inset to fig .
[ tffig](b ) and suggests that there is a scaling relation between these two parameters .
although this pair of @xmath1-phase points lie rather close , a simple linear scaling @xcite would be reasonably consistent with the data .
more generally , a @xmath61 scaling has been suggested to apply when a broader range of molecular superconductors is considered @xcite , but more data would be required to test whether this holds or not within the subset of @xmath1-phase bedt - ttf compounds .
we now turn to the local magnetic fluctuations in the normal state of this system , which we probe with lf @xmath0sr measurements made in a field of @xmath62 mt . in these measurements
the initial muon spin is directed parallel to the applied field and we measure the polarization @xmath63 along the same direction via the muon asymmetry function @xmath64 $ ] .
dynamics in the local magnetic field distribution will cause muon spins to flip leading to a depolarization with relaxation rate @xmath65 .
the small applied longitudinal field was intended to quench the contribution to the spectra of background from nuclear moments allowing the contribution due to electron spin dynamics to be discerned . a typical lf @xmath0sr spectrum measured at 20 k is shown in fig . [ lambdat](a ) .
the relaxing amplitude is quite small , but this is not unexpected given previous zero field muon results measured on a related bedt - ttf material @xcite . in that study , muons were shown to be sensitive to magnetic order within the organic layers but only a small relaxing asymmetry was observed .
we therefore expect that our results will reflect magnetic fluctuations within the organic layers of @xmath1-(bedt - ttf)@xmath5 . above 100
k our spectra are best described with a single exponential function with relaxation rate @xmath65 characteristic of dynamic fluctuations in the local magnetic field at the muon sites in the material . for @xmath66
k , there is a single , highly damped , temperature dependent oscillation visible for times @xmath67s . in this context ,
spectra of this form suggest a contribution from quasistatic disordered magnetic moments which is typically described by a kubo - toyabe ( kt ) function@xcite @xmath68 where the data are best fitted with a constant field width @xmath69 mhz for all @xmath66 k. we note that this is quite large for randomized nuclear fields ( where typically @xmath70 mhz ) and given the small size of the electronic moments observed with muons in bedt - ttf layers previously @xcite it may be that this signal results from a population of static disordered electronic moments .
the origin of such quasistatic electronic moments is unclear although we note that there is no decrease in @xmath71 with increasing @xmath72 as would occur in the presence of magnetic order .
we note that it is possible that this signal arises from those portions of the sample not contributing to the superconducting vl in tf measurements .
the dynamic contribution to the spectra for @xmath73 k was found to be best modelled by multiplying the kt function by an exponential factor @xmath74 and the data were fitted to the resulting functional form @xmath75 the extracted relaxation rates @xmath65 are shown in fig .
[ master](a ) and inset in fig .
[ lambdat](b ) .
other parameterizations are possible but lead to less successful fits . however , for comparison , a fit across the measured temperature regime of a single exponential with relaxation rate @xmath65 is shown in fig .
[ master](a ) , where we see that both track the same behavior . at temperatures
@xmath66 k , the dominant trend is a roughly linear increase in relaxation rate with @xmath72 up to around 100 k , above which it drops precipitously [ inset fig . [ lambdat](b ) ] .
the linear regime observed for @xmath66 k is typical of a korringa law relaxation @xcite , resulting from flip - flop transitions of electronic and muon spins , and which is expected to lead to a relaxation rate @xmath76 , where @xmath77 is the imaginary part of the dynamic magnetic susceptibility and @xmath78 is the hyperfine coupling .
for the case of a @xmath79-independent coupling , we obtain @xmath80^{2}$ ] , showing that the muon relaxation rate probes the local spin susceptibility @xmath81 . the sharp decrease in relaxation rate around 100 k
broadly coincides with region i , which may suggest a link with the ethylene disorder effects and glass - like transition in this region .
given the korringa law behavior observed at low temperatures , it is likely that the crossover to nonmetallic behavior of the system near region i [ fig .
[ master](b ) ] changes the nature of the muons coupling to the material and that at these elevated temperatures a different , dominant relaxation channel presumably opens up .
to further investigate the temperature evolution of the local susceptibility from the korringa relaxation we plot @xmath82 against @xmath72 in fig .
[ lambdat](b ) .
we see that on warming above @xmath4 the quantity @xmath83 peaks at around 15 k before decreasing sharply around region iii ( @xmath84 k ) indicating a significant decrease of the local magnetic susceptibility in this region . above 30
k @xmath83 decreases only gradually throughout region ii until dropping off suddenly in region i. our muon @xmath83 results are quite different to the nmr @xmath19 results [ fig.[master ] ] where a parger peak is seen in region ii .
we do not see any dramatic change of behavior in region ii in the @xmath0sr results , while there is no suggestion of any anomaly in the nmr behavior in regions i and iii .
in fact , the nmr results for @xmath85 k@xmath86 k are well described by a model of antiferromagnetic spin fluctuations@xcite based on the mmp model@xcite , which predicts @xmath87 where @xmath88 and @xmath31 are constants that depend on the correlation length of the spin fluctuations whose energy scale is determined by the temperature parameter @xmath89 . as shown fig . [
lambdat](b ) the model fits the muon @xmath65 only passably well for @xmath66 k becomes gradually worse as region iii approached from above . however , given that the muon sites in this material are likely to be different to the @xmath90c site where the nuclear resonance is achieved@xcite , it is possible that the muon will probe a quite different field distribution to that probed in nmr .
we note further that our muon results are also quite different to the lf @xmath0sr behavior observed in molecular superconductors such as the alkali - fullerides @xcite where @xmath83 is quite featureless above @xmath4 .
in conclusion , our investigation of the superconducting and normal state properties of @xmath1-(bedt - ttf)@xmath2cu[n(cn)@xmath2]br using implanted muons has allowed us to identify a number of experimental features of this system . in the superconducting phase
two important effects are seen : firstly , a reduced superconducting signal fraction that is affected by sample cooling rate and is consistent with the phase separation suggested by earlier studies and secondly , a vortex transition taking place between 40 mt and 80 mt , that reduces the width of the magnetic field distribution for higher measurement fields . by taking these two effects carefully into account
we have obtained an improved and more reliable estimate for the @xmath91 in - plane penetration depth @xmath47=0.47(1 ) @xmath22 m and find that the trend of @xmath41 increasing with superfluid stiffness @xmath92 for @xmath1-(bedt - ttf)@xmath93x superconductors is consistent with the overall trend for molecular superconductors , whereas previous data had suggested that @xmath1-(bedt - ttf)@xmath2cu[n(cn)@xmath2]br was anomalous in this regard . in the normal state
we find a large peak in the longitudinal muon spin relaxation rate around 100 k , coinciding with the region where ethylene disorder effects are prevalent .
we also observe an enhancement of the local magnetic spin susceptibility above @xmath4 where vortex fluctuations occur in nernst measurements .
it would be desirable in future to extend @xmath0sr studies of the normal state magnetic properties to the other members of this series in order to follow how these features evolve across the phase diagram .
we are grateful to alex amato , andrew steele and peter baker for experimental assistance and to epsrc ( uk ) for financial support .
part of this work was performed at s@xmath22s and part at the stfc isis facility and we are grateful to psi and stfc for the provision of beamtime . work supported by u. chicago argonne , llc , operator of argonne national laboratory ( argonne " ) .
argonne , a u.s .
department of energy office of science laboratory , is operated under contract no .
de - ac02 - 06ch11357 .
luke , b.j .
sternlieb , w.d .
uemura , j.h .
brewer , t.m .
riseman , c.e .
stronach , g. saito , h. yamochi , h.h .
wang , a.m. kini , k.d .
carlson , j.m .
williams , phys .
lett * 68 * , 1923 ( 1992 ) .
harshman , a.t .
fiory , r.c .
haddon , m.l .
kaplan , t. pfiz , e. koster , i. shinkoda and d.ll .
williams , phys .
b * 49 * , 12990 ( 1994 ) . r.s .
hayano , y. j. uemura , j. imazato , n. nishida , t. yamazaki and r. kubo , phys .
b * 20 * , 850 ( 1979 ) . | a muon - spin relaxation ( @xmath0sr ) investigation is presented for the molecular superconductor @xmath1-(bedt - ttf)@xmath2cu[n(cn)@xmath2br ] .
evidence is found for low - temperature phase - separation , with only a fraction of the sample showing a superconducting signal , even for slow cooling .
rapid cooling reduces the superconducting fraction still further . for the superconducting phase
, the in - plane penetration depth is measured to be @xmath3 m and evidence is seen for a vortex decoupling transition in applied fields above 40 mt .
the magnetic fluctuations in the normal state produce korringa behavior of the muon spin relaxation rate below 100 k , a precipitous drop in relaxation rate is seen at higher temperatures and an enhanced local spin susceptibility occurs just above @xmath4 . |
the origin of parity violation can be explained within the frame work of the standard model ( sm ) of particle physics through the weak interactions , but there is no clear understanding of the origin of time - reversal violation in nature .
the presence of a non - zero electric dipole moment of a non - degenerate physical system is a direct signature of parity ( @xmath1 ) and time - reversal ( @xmath2 ) symmetry violations @xcite . in this paper
we present a study of atomic edms of closed - shell atoms which arise mainly from the tensor - pseudo tensor ( t - pt ) electron - nuclear interactions and the nuclear schiff moment ( nsm ) . at the elementary particle level ,
the origin of closed - shell atomic edms is attributed to the @xmath1 and @xmath2 violating electron - quark interactions and quark - quark interactions which are predicted by the lepto - quark models , susy , etc @xcite .
the limits on the t - pt coupling constant ( @xmath3 ) has been obtained from the comparison of the most recent experimental result of @xmath0 atomic edm @xcite , @xmath4 and the enhancement factor ( ratio of atomic edm to the coupling constant of the interaction in question ) calculated in @xcite .
the calculation uses coupled - perturbed hartree - fock ( cphf)@xcite theory and the calculated value of the atomic edm is @xmath5 combining with experimental result , a limit on @xmath3 can be derived , @xmath6 where @xmath7 is the nuclear spin .
a non - zero value of @xmath3 would imply physics beyond the standard model .
improving the accuracy of the calculations would lead to a more stringent bound on @xmath3 .
the cphf theory accounts for two - particle two - hole type of electron correlations to all orders in perturbation .
in addition to this , for accurate atomic calculations@xcite , it is important to include other correlation effects which are absent in the cphf theory .
coupled - cluster ( cc ) theory @xcite is one of the many - body methods which is non - perturbative and incorporates all forms of correlation effects @xcite .
cc theory with all levels of excitations is equivalent to all order many - body perturbation theory .
it has been successfully applied for calculation of a variety of atomic properties @xcite .
it has also been applied to atomic hg for calculating transition energies @xcite .
the nsm(denoted by operator @xmath8 ) can originate from the nucleon - nucleon interactions or a nucleon edm . at the elementary particle level it can arise from the interaction between the quarks and the chromo edm of the quarks .
the coupling constants associated with these interactions can be predicted by multi - higgs , susy @xcite etc .
the dependence of the t - pt and nsm interactions on the nuclear spin makes closed - shell atoms , in particular , those having non - zero nuclear spin the best candidates to look for edms sensitive to the nuclear sector .
for @xmath0 , the edm induced by the nsm is calculated and parameterized in terms of the schiff moment operator @xmath8 .
the most recent calculation @xcite gives @xmath9 according to the recent work @xcite the expression for the p and t - odd nuclear potential arising from the nuclear schiff moment @xmath10 where @xmath11 and @xmath12 is the electron coordinate and @xmath13 is the nuclear density .
the interaction hamiltonian of electrons with this potential is @xmath14 where @xmath15 is the charge of the electron . if @xmath13 is the normalized density function , which is 1 for @xmath16 , where @xmath17 is the nuclear radius and @xmath18 is a small number .
the single electron matrix elements of @xmath19 are @xmath20 \frac{\rho(r)}{b}rdr .\ ] ] where @xmath21 and @xmath22 are the large and small components of the relativistic dirac - wavefunctions @xmath23 .
the matrix elements of @xmath19 between the states @xmath24 and @xmath25 is identical to the above expression .
the starting point of obtaining the coupled - cluster equations is the relativistic atomic hamiltonian in the dirac - coulomb approximation in atomic units ( @xmath26 , @xmath27 and @xmath28 ) @xmath29 + \sum_{i < j}^n \frac{1}{r_{ij } } , \label{eqdch}\ ] ] where c is velocity of light , @xmath30 and @xmath31 are the dirac matrices , @xmath32 is the coulomb potential energy between two electrons , and @xmath33 is the nuclear potential . in the above hamiltonian , the rest mass energy is subtracted from the total energy eigen values .
this is the hamiltonian of an atomic system considering only the electrostatic interactions .
the single particle equations are obtained by approximating the two - electron term in eq.(@xmath34 ) by a central field potential @xmath35 , known as the dirac - fock potential @xcite , then @xmath36 + v_{\rm es}. \label{eqdc}\ ] ] the residual coulomb interaction @xmath37 the non - central ( or ) correlation effects are included by treating @xmath38 as a perturbation .
the single electron wavefunctions @xmath39 satisfy the schroedinger equation @xcite @xmath40|\psi_a\rangle= \epsilon_a |\psi_a\rangle \label{dfeqn}\end{aligned}\ ] ] and are expressed in the relativistic framework in terms of the two - component form disucssed in succeeding sections .
the eq.(@xmath41 ) is obtained by variational extremization @xcite .
the perturbed equations are obtained by introducing the @xmath1 and @xmath2 violating tensor - pseudo tensor interaction hamiltonian in addition to the residual coulomb interaction .
the exact atomic wavefunction in cc formalism @xmath42 , where @xmath43 is the reference state , @xmath44 is the cluster operator . in our calculations
, we use the singles and doubles approximation @xmath45 , the second quantized form of the operators are @xmath46 the many - body schroedinger equation of the dirac - coulomb hamiltonian for an atomic system , in a state @xmath47 in the cc formalism @xmath48 operating from left side by @xmath49 @xmath50 expressing @xmath51 in normal ordered form , @xmath52 , where @xmath53 is the dirac - fock energy .
then , @xmath54 projecting eq.(@xmath55 ) with singly and doubly excited states @xmath56 and @xmath57 respectively , the single and double excitation cluster amplitude equations are @xcite @xmath58 the @xmath59 operator is the general edm operator . in the present paper , @xmath59 denotes @xmath19 , which is a single - particle operator .
consider the nsm perturbed schroedinger equation @xmath60 where @xmath61 and @xmath62 .
taking upto one order in @xmath63 in the exponent , @xmath64 substituting for @xmath65 in the above equation and comparing zeroth and first order terms of @xmath66 on both sides , @xmath67 and @xmath68 multiplying eq.(@xmath69 ) by @xmath70 on both sides and substitute @xmath71 @xmath72 and eq.(@xmath73 ) can be written as @xmath74 since @xmath75 and @xmath70 commute and @xmath76 is the correlation energy .
operate eq.(@xmath77 ) on both sides by @xmath78 @xmath79 operating by @xmath78 on eq.@xmath80 @xmath81 subtracting eq.(@xmath82 ) from eq.(@xmath83 ) @xmath84 |\phi_0\rangle = -{\overline h}_{\rm edm } |\phi_0\rangle , \label{beq}\ ] ] where @xmath85 and @xmath86 is any operator .
the equation for the @xmath59 perturbed singles and doubles cluster amplitudes can be derived from the basic equation , eq.(@xmath87 ) by projecting on both sides of the equation with singly and doubly excited determinantal states .
@xmath88|\phi_0\rangle & = & -\langle\phi_a^r|\overline h_{\rm edm}|\phi_0\rangle \nonumber \\ \langle\phi_{ab}^{rs}|\left[\overline h_n , t^{(1)}\right]|\phi_0\rangle & = & -\langle\phi_{ab}^{rs}|\overline h_{\rm edm}|\phi_0\rangle \end{aligned}\ ] ] which are equivalent to @xmath89 further expanding @xmath90
, these equations can be cast in the form of a system of linear matrix equations @xmath91 where @xmath92 , @xmath93 , @xmath94 , @xmath95 are the sub blocks of the dressed @xmath96 matrix elements and @xmath97 and @xmath98 are the sub blocks of the dressed @xmath59 matrix elements .
the edm of the atom in the state @xmath99 is @xmath100 where d is the electric dipole operator . substituting the above expression for @xmath101 and keeping terms only of order @xmath66
, we obtain the expression for edm ( the symbol of contraction appears using @xmath102 , where the curly brackets refer to normal ordering and the expectation value of normal ordered operator between the vacuum states is zero ) @xmath103|\phi_0\rangle } { \langle \psi_0|\psi_0\rangle } , \ ] ] which can also be written as @xmath104|\phi_0\rangle } { \langle \psi_0|\psi_0\rangle } , \label{edmeqn1}\ ] ] where @xmath105 . using the fact that @xmath70 and @xmath106 operators are odd and @xmath75 operator is even under parity
, the bra and the ket vectors in the above expression must have the same parity . to simplify the calculations ,
we expand @xmath107 as @xcite @xmath108 in eq.(@xmath109 ) , the one - body nature of the electric dipole operator @xmath106 restricts the maximum possible contractions with @xmath75 to just two .
define @xmath110 then @xmath111 the last step follows as the two terms are the complex conjugates of each other and give equal contributions .
substituting the expanded form of @xmath112 @xmath113 |\phi_0\rangle/\langle \psi_0|\psi_0\rangle \nonumber \\ & = & 2 \left [ \langle \phi_1|d e^{t^{(0)}}|\phi_0\rangle + \langle \phi_1|\sum_{n=1}^{\infty } \frac{1}{n ! } \left ( { t^{(0)}}^\dagger\right ) ^n d e^{t^{(0)}}|\phi_0\rangle \right ] /\langle
\psi_0|\psi_0\rangle \label{eqn - final}\end{aligned}\ ] ] the complexity of the above expression can be mitigated by exploiting the fact that not all the terms containing the @xmath114 operators contribute to the infinite summation . in this scheme ,
the zeroth order @xmath115 consists of terms without @xmath114 operator , first order @xmath115 consists of terms having one order of @xmath114 , and so on .
the unlinked terms of the numerator cancel with the denominator and only linked terms contribute in the numerator . at the linear level ,
@xmath116 following are the terms contributing to the edm expectation value at the linear level : @xmath117 the two terms in eq.@xmath118 are complex conjugates of each other .
hence @xmath119 substituting for @xmath120 the expression for the atomic edm @xmath121 |\phi_0\rangle \label{atom_sample}\ ] ]
the single particle orbitals for all calculations in the subsequent sections were generated using the gaussian basis @xcite set expansion whose salient features are presented in this section . in the central field approximation ,
the solution of the dirac equation in terms of the four component spinors is given by @xmath122 where @xmath123 and @xmath124 are the large and the small components of the radial wavefunctions expanded in terms of the basis sets@xcite , @xmath125 where the summation over the index @xmath126 runs over the number basis functions @xmath127 , @xmath128 and @xmath129 correspond to the large and small components and @xmath130 and @xmath131 are their expansion coefficients for each value of @xmath132 .
the functions @xmath128 are chosen to be the two parameter gaussian type orbitals ( gtos ) @xcite and large and small components are related by the condition of kinetic balance @xcite .
the calculations presented in this paper correspond to the even tempered ( et ) basis set . for the nuclear density we use the two parameter fermi distribution @xmath133 where @xmath134 is the radial coordinate , @xmath135 is the skin thickness parameter and @xmath136 , the half - charge radius @xcite . in the expression of @xmath137 , the @xmath138 is the nuclear mean - square radius . the quantity @xmath139 is determined by normalising @xmath140 over a spherical volume .
the radial grid has the form , @xmath141 , @xmath142 is the total number of grid points . .[gauss - table]no . of basis functions used to generate the even tempered dirac - fock orbitals and the corresponding value of @xmath143 and @xmath31 used .
the total number of active orbitals are shown in the brackets for _ _ a__ctive holes . [ cols= " <
, < , < , < , < , < , < , < , < , < " , ] limits on the p and t violating parameters at the elementary particle level can be derived from the tensor - pseudo tensor coupling constant ( @xmath3 ) and the nuclear schiff moment ( @xmath144)@xcite .
the results of various p and t violating parameters are shown in table.@xmath145 .
a detailed discussion on the derivation of these parameters has been presented in @xcite . using these important parameters
, the observable edms can be connected to the p and t violating coupling constants predicted by the underlying elementary particle theories .
the bounds on these coupling constants can further be used to constrain various models of elementary particle physics .
l l p and t violating parameter at the & limits obtained from elementary particle level & linear ccedm @xmath146 & @xmath147 @xmath148 & @xmath149 @xmath150 & @xmath151 e@xmath152 & @xmath153 e cm
13 commins e d 1999 _ advances in atomic , molecular and optical physics _ * 40 * 1 - 56 .
lee t d and yang c n 1957 brookhaven national laboratory report no .
bnl 443-t91 .
landau l d 1957 nucl.phys . * 3 * 127 .
barr s m and marciano w j 1989 _ cp violation _
, ed c jarlkog ( world scientific , singapore ) .
romalis m v , griffith w c , jacobs j p and fortson e n 2001 _ phys .
rev . lett._*86 * 2505 - 08 .
jacobs j p , klipstein w m , lamoreaux s k , heckel b r and fortson e n 1993 _ phys .
lett . _ * 71 * 3782 - 85 .
ann - marie martensson - pendrill 1985 _ phys .
lett _ * 54 * 1153 - 55 .
rossky p j and karplus m 1980 _ j. chem .
phys . _ * 73 * 6196 .
caves t c and karplus m 1969 _ j. chem .
phys . _ * 50 * 3649 .
liu z w and hugh p kelly , 1992 _ phys.rev.a . _ * 45 * , r4210-r4213 .
coester f 1958 _ nucl .
phys . _ * 7 * 421 - 4 ; coester f and kmmel h 1960 _ nucl
* 17 * 477 - 85 ; ek j 1966 _ j. chem .
* 45 * 4256 - 66 .
bartlett r j 1995 _ modern electronic structure theory _ ,
iii , 1047 _ edited by _
d r yarkony , published by world scientific , singapore .
ephraim eliav , uzi kaldor and yasuyuki ishikawa 1995 _ phys . rev .
a _ * 52 * , 2765 - 2769 .
rupsi pal , safronova m s , johnson w r , andrei derevianko , and sergey g. porsev 2007 _ phys . rev .
a _ * 75 * , 042515(1)-042515(11 ) .
rodney j. bartlett and monika musia 2007 _ rev .
_ * 79 * , 291 - 352 .
merlitz h , geetha gopakumar , rajat k chaudhuri , das b p , uttam sinha mahapatra , and debashis mukherjee 2001 _ phys .
a _ * 63 * , 022507(1 ) - 022507(4 ) yong - ki kim 1967 _ phys .
rev . _ * 154 * , 17 - 39 .
falk t , olive k a , pospelov m and roiban r , 1999 _ nucl.phys._b * 560 * 3 - 22 .
dzuba v a , flambaum v v , ginges j s m and kozlov m g , 2002 _ phys.rev._a * 66 * 012111 ( 7 pages ) .
ginges j s m and flambaum v v 2004 _ physics reports _ * 397 * 63 - 154 .
flambaum v v and ginges j s m 2002 _ phys . rev._a * 65 * 032113 ( 9 pages ) .
lindgren i and morrison j 1985 _ atomic many - body theory _
( berlin : springer - verlag ) .
szabo a and ostlund n s _ modern quantum chemistry _
( new york : dover publications ) .
angom d and latha k v p , 2006 _ calculation of atomic electric dipole moment of closed - shell atoms using coupled - cluster method _
( unpublished , http://www.iiap.res.in/research/napp/main.html ) .
latha k v p , 2006 _ many - body theory of electric dipole moments of atoms and it s implications for the standard model of particle physics _
( phd thesis submitted to mangalore university , india ) .
mohanty a k and clementi e , 1989 _ chemical physics letters _ , * 157 * , 348 - 352 .
stanton r e and havriliak s , 1984 _ journal of chemical physics _ ,
* 81 * , 1910 - 18 .
grant i p and quiney h m , 1988 _ advances in atomic and molecular physics _ * 23 * , 37 .
parpia f a and mohanty a k , 1992 _ phys .
rev._a * 46 * 3735 - 45 .
geetha k p 2001 _ relativistic many - body studies of parity non - conservation in heavy atomic ions _ , ph d thesis , bangalore university . | the electric dipole moments of closed - shell atoms are sensitive to the parity and time - reversal violating phenomena in the nucleus .
the nuclear schiff moment is one such property , it arises from the parity and time reversal violating quark - quark interactions and the quark - chromo electric dipole moments .
we calculate the electric dipole moment of atomic @xmath0 arising from the nuclear schiff moment using the relativistic coupled - cluster theory .
this is the most accurate calculation of the quantity to date . our calculations in combination with the experiment data provide important insights to the p and t violating coupling constants at the elementary particle level .
in addition , a new limit on the tensor - pseudo tensor induced atomic edm , calculated using the relativistic coupled - cluster theory is also presented . |
28 b. r. mollow 1969 , phys . rev . * 188 * , 1969 ; b. r. mollow 1972 , phys .
rev . a * 5 * , 2217 ; f.y.wu et .
1977 , phys .
lett . * 38 * , 1077 .
a. suguna and g. s. agarwal 1979 , phys .
a * 20 * , 2022 ; r. w. boyd et al .
1981 , _ ibid . _ * 24 * , 411 ; r. s. bennink et .
al . 2001 , _ ibid . _ * 63 * , 033804 ; m. e. crenshaw and c. m. bowden 1991 , phys .
lett . * 67 * , 1226 . s. e. harris , j. e. field , and a. imamoglu 1990 , phys . rev .
64 , 1107 ; s. e. harris 1997 , phys . today *
50*(7 ) , 36 . o. a. kocharovskaya and y. i. khanin 1988 , jetp
. lett . * 48 * , 630 ; s. e. harris 1989 , phys .
lett . * 62 * , 1033 ; m. o. scully , s .- y .
zhu , and a. gavrielides 1989 , _ ibid . _ * 62 * , 2813 ; a. s. zibrov et .
al . 1995 , _ ibid . _ * 75 * , 1499 . m. o.
scully 1991 , phys .
lett . * 67 * , 1855 ; m. fleischhauer et .
1992 , phys .
a * 46 * , 1468 ; h. friedmann and a. d. wilson - gordon 1993 , opt . commun . * 98 * , 303 ; o. kocharovskaya , p. mandel , and m. o. scully 1995 , phys
. rev . lett . * 74 * , 2451
. t. quang , h. freedhoff 1993 , phys .
a * 48 * , 3216 ; c. szymanowski and c. h. keitel 1994 , j. phys .
b : at . mol .
* 27 * , 5795 ; u. akram , m. r. b. wahiddin , z. ficek 1998 , phys .
a * 238 * , 117 ; m. haas , c. h. keitel 2003 , optics commun . *
216 * , 385 ; g. li et .
2000 , j. phys .
b * 33 * , 3743 . o. kocharovskaya , y. rostovtsev , m. o. scully 2001 , phys
lett . * 86 * , 628 ; d. f. philips et .
al . 2001 , _ ibid . _ * 86 * , 783 .
peng zhou and s. swain 1996 , phys .
lett . * 77 * , 3995 ; c. h. keitel 1999 , _ ibid . _
* 83 * , 1307 .
s. e. harris 1996 , phys .
lett . * 77 * , 5357 . g. s. agarwal , robert w. boyd 1999 , phys .
rev . a * 60 * , r2681 .
a. s. manka et al .
1994 , phys .
lett . * 73 * , 1789 .
m. macovei , j. evers , and c. h. keitel 2003 , phys .
91 * , 233601 ; m. macovei , j. evers , and c. h. keitel 2005 , phys . rev .
a * 71 * , 033802 .
a. wicht , r .- h .
rinkleff , l. spani molella , and k. danzmann 2002 , phys .
a 66 , 063815 .
a. v. andreev , v. i. emelyanov , and yu .
a. ilinskii , _ cooperative effects in optics .
superfluorescence and phase transitions _
( iop publishing , london , 1993 ) .
r. r. puri , _ mathematical methods of quantum optics _ ( springer , berlin 2001 ) .
r. h. dicke 1954 , phys . rev . * 93 * , 99 .
bachor and t. c. ralph , _ a guide to experiments in quantum optics _ ( wiley - vch , weinheim 2004 )
. m. o. scully and m. s. zubairy , _ quantum optics _ , cambridge university press ( 1997 ) .
r. bonifacio , dae m. kim and marlan o. scully 1969 , phys . rev . * 187 * , 441 ; r. gilmore , c. m. bowden and l. m. narducci 1975 , phys . rev .
a * 12 * , 1019 ; g. s. agarwal et .
1979 , phys .
* 42 * , 1260 . | the collective interaction via the environmental vacuum is investigated for a mixture of two deviating multi - atom ensembles in a moderately intense laser field . due to the numerous inter - atomic couplings
, the laser - dressed system may react sensitively and rapidly with respect to changes in the atomic and laser parameters .
we show for weak probe fields that in the absence of absorption both the index of refraction and the group velocity may be modified strongly and rapidly due to the collectivity .
the presence of strong laser fields are known to substantially modify the absorptive and dispersive properties of atomic samples @xcite .
the numerous effects put forward already for single atoms include the splitting of spectral lines @xcite , electromagnetically induced transparency @xcite , lasing without population inversion @xcite , enhanced indeces of refraction @xcite , stopping of light @xcite and ultra - narrow spectral lines @xcite . for larger ensembles and not too low densities , collective phenomena were further pointed out to drastically affect laser - driven media @xcite . in particular , collectivity due to nonlinearities in a plasma may render the medium transparent @xcite , while laser - mediated local effects arising from dipole - dipole interactions between atoms may alter the appearance of band gaps in optically dense materials @xcite and induce piezophotonic switching @xcite .
efficient schemes to control the collective quantum dynamics in general were further demonstrated @xcite via employing interferences and fast switching schemes among collective atomic dressed - states .
particular interest , however , is in the rapid control of dispersive properties of collective systems , e.g. for quantum gates or high precision measurements @xcite . in this letter
we show that the mutual interactions of atoms via the quantum fluctuations of the surrounding electromagnetic field ( emf ) @xcite are suitable to generate transparent media with large indeces of refraction of order 10 or high dispersion of arbitrary sign .
in particular we point out that the effects may be considerably larger than for the same amount of independent atoms and may be set up on a time scale of 10@xmath0s ( ghz ) .
furthermore , the group velocity of a weak probe field propagating through the strongly driven two - level medium may be substantially slowed down or strongly accelerated . for this purpose
, we consider an atomic system consisting of two ensembles of two - level atoms , numbered @xmath1 and @xmath2 , with densities of order @xmath3 , and with somewhat different transition frequencies , and interacting with a single moderately strong laser field ( see fig .
[ fig-1 ] ) .
the corresponding rabi frequencies are @xmath4 and spontaneous decay of all closely spaced atoms occurs via interaction with a common electromagnetic field reservoir with rates @xmath5 from excited states @xmath6 , respectively . in order to treat the atoms
uniformly we suppose that @xmath7 where @xmath8 , @xmath9 and @xmath10 are the largest dimension of the sample , the light velocity and the collective decay time , respectively . when scanning the composed atomic sample with , say for instance , a dye or diode laser @xcite , and depending on the resonance condition for each kind of atom and thus the employed laser frequency , generally one or the other atom specy may dominate the final steady - state collective behavior . schematic diagram depicting means of control over refractive properties of two laser - driven two - level atomic ensembles . (
a ) the involved energy levels of the two two - level atomic systems and rabi frequencies are denoted with @xmath11 and @xmath12 .
( b ) the index of refraction and thus the deviation angle @xmath13 of a weak probe field ( @xmath14 ) may be substantially and rapidly manipulated by an applied laser field ( @xmath15).,width=291 ] in the conventional mean - field and rotating - wave approximation , the interaction of the laser - driven atomic sample with the surrounding emf bath is described in a frame rotating with the laser frequency @xmath15 , via the hamiltonian : @xmath16 where @xmath17 , @xmath18 = @xmath19 $ ] and @xmath20=@xmath21 here the first and the second terms in eq .
( [ hm ] ) represent the free emf ( @xmath22 ) and the free atomic plus laser - atom interaction hamiltonian ( @xmath18 ) , respectively .
the last term of eq .
( [ hm ] ) describes the interaction ( @xmath20 ) of the atoms with the environmental vacuum modes ( @xmath23 ) . the collective atomic operators @xmath24 @xmath25 @xmath26 = @xmath27^{\dagger}$ ] , @xmath28 = @xmath29 @xmath30 satisfy the standard commutation relations for quasispin operators , i.e. , @xmath31 $ ] = @xmath32 , @xmath33 $ ] = 2@xmath34 , @xmath35 .
@xmath36 = @xmath37 denote the detuning of the atomic transitions frequencies @xmath38 to the laser frequency , and @xmath39 .
@xmath40 corresponds to the transition dipole matrix elements of the atoms .
@xmath41 where @xmath42 is the photon polarization vector while @xmath43 is the emf quantization volume .
the master equation corresponding to the hamiltonian ( [ hm ] ) in the born - markov approximation then reads @xmath44 = -\sum_{i , j \in \{a , b\}}\bigl ( \sqrt{\gamma_{i}\gamma_{j}}[s^{(i)}_{+},s^{(j)}_{-}\rho ] \nonumber \\ & + & \sqrt{r_{i}r_{j}}[s^{(i)}_{z},s^{(j)}_{z}\rho ] \bigr ) + h.c . \label{meq}\end{aligned}\ ] ] the diagonal ( @xmath45 ) contribution of the first term on the right - hand side describes the collective damping due to the spontaneous emission of atoms , while that proportional to @xmath46 involves the mutual exchange of photons among the different types of atoms in the sample and is very sensitive relative to the splitting frequency @xmath47 = @xmath48 - @xmath49 with @xmath50 . the dipole - dipole interactions @xmath51 among the emitters are omitted here , an approximation valid as soon as @xmath52 . the last term of the master equation eq .
( [ meq ] ) involving the collision rates @xmath53 accounts for collisional damping of atoms which alter the phase of the atomic state but not its population @xcite . in the intense - field limit ( @xmath54 ) , the master equation eq .
( [ meq ] ) transforms into the dressed - state picture ( @xmath55 for @xmath56 , @xmath57 ) via @xmath58 = @xmath59 , @xmath56 . in the secular approximation ,
i.e. upon omission of the terms oscillating with rabi frequency @xmath60 and larger , and without the cross - damping contribution ( @xmath61 ) , the resulting dressed state master equation results in the exact steady - state solution of the form @xmath62 here @xmath63/[\gamma_{i}\sin^{4}{\theta_{i}}+r_{i}\sin^{2}(2\theta_{i})/4 ] \bigr)$ ] , @xmath64 = @xmath65 - @xmath66 , @xmath67 while @xmath68 is chosen such that @xmath69 the influence of the cross - damping terms with respect to the final collective steady - state dynamics can be estimated approximately for larger samples , while for single - atom systems ( @xmath70 ) this can be carried out exactly by solving the respective equations of motion for the atomic variables . in the dressed master equation , the sideband contribution proportional to @xmath71 oscillates at the relative frequency @xmath72 which can be nonzero as , in general , the atoms are subject to different detunings and laser intensities . if @xmath73 such contributions can be omitted in the secular approximation ( as well as for smaller samples ) , and the solution in eq .
( [ msol ] ) is then applicable .
when @xmath74 and the strong laser field being detuned far off the frequency range @xmath75 , i.e. @xmath76 , then eq . ( [ msol ] ) can be employed with @xmath77 .
assuming @xmath78 , and @xmath79 , the steady - state solution can further be obtained from eq .
( [ msol ] ) in the limit @xmath80 . in
what follows , however , we neglect the cross - damping contributions and thus restrict ourselves to the case @xmath81 . on employing the atomic coherent states for two - level atoms @xcite and the solution in eq .
( [ msol ] ) , we derive the expectation values for any collective atomic correlators of interest .
in particular , the steady - state expectation values for the collective dressed - state inversion operators @xmath82 can be obtained for @xmath83 and with @xmath84 : @xmath85 in what follows we concentrate on the case with almost equal parameters for the two types of atoms , i.e. @xmath86 , @xmath87 , @xmath88 and @xmath89 .
fig.([fig-2ab ] ) depicts the steady - state dependence of the dressed - state inversion operators as a function of the ratio @xmath90 .
as the atomic transition frequencies for the atoms of type @xmath11 and @xmath12 differ from each other , the steady - state collective populations behave differently as well .
note that , due to collectivity , the collective dressed state populations may be transferred abruptly and rapidly from one dressed state to another as the laser detuning @xmath36 changes its sign .
a few- atom system is less sensitive relative to the laser detuning in the sense of fast switching .
thus , at this particular point , @xmath91 with @xmath92 one may switch the absorption properties of a weak probe field from positive to negative gain ( or vice versa ) while the dispersive features are strongly enhanced . for a pencil - shaped sample with length @xmath93 , transversal area @xmath94 , @xmath95 , @xmath96 , and
@xmath97 we estimate a switching time @xmath98 of about @xmath99 .
the secular approximation can be applied here if @xmath100 , as @xmath101 we proceed by calculating the refractive properties of a very weak field probing the strongly driven atomic sample .
the linear susceptibility @xmath102 of the probe field , at frequency @xmath14 , can be represented in terms of the fourier transform of the average value of the two - time commutator of the atomic operator as @xmath103\rangle_{s}. \label{chi}\end{aligned}\ ] ] note that the steady - state ( subindex @xmath104 ) of the atomic correlators in eq .
( [ chi ] ) should be calculated with the help of eqs .
( [ drs],[msol ] ) . introducing eq .
( [ drs ] ) in eq .
( [ chi ] ) , and making use of both the secular approximation and the quantum regression theorem @xcite , together with eq .
( [ msol ] ) , then the dispersion and absorption features can be described via : @xmath105 , \nonumber \\
\chi^{''}(\delta_{p } ) & = & \sum_{i \in \{a , b\}}\frac{\bar n_{i}d^{2}_{i}}{\gamma_{i}\hbar}\frac{\langle r^{(i)}_{z}\rangle_{s}}{n_{i } } \bigl [ \sin^{4}{\theta_{i}}\frac{\tilde \gamma_{i}}{\tilde \gamma^{2}_{i}+(\tilde \delta^{(i)}_{p } + 2\bar \omega_{i})^{2 } } \nonumber \\ & -&\cos^{4}{\theta_{i}}\frac{\tilde \gamma_{i}}{\tilde \gamma^{2}_{i}+(\tilde \delta^{(i)}_{p } - 2\bar \omega_{i})^{2 } } \bigr ] .
\label{refr}\end{aligned}\ ] ] here @xmath106 , while @xmath107 is the atomic density .
@xmath108 = @xmath109 corresponds to the detuning of the weak probe field frequency with respect to the driving field , while @xmath110 describes the non - diagonal collective damping with @xmath111 $ ] + @xmath112 $ ] and @xmath113 respectively .
it should further be noted that in eqs .
( [ refr ] ) we have employed the so - called decoupling scheme for symmetrical atomic correlators as valid for @xmath114 @xcite . in the absence of collective effects ( @xmath115 ) , eqs .
( [ refr ] ) reduces , as it should , to the correct result for @xmath1 and @xmath2 independent atoms . on inspecting eqs .
( [ refr ] ) ( involving @xmath82 ) and fig .
( [ fig-2ab]a ) one can easily recognize that the susceptibility @xmath116 is substantially enhanced via collective effects .
figures ( [ fig-3ab ] ) depict the steady - state dependence of the linear susceptibility with respect to the strong laser detunings while keeping fixed probe - field frequencies .
strong gain , strong positive or negative dispersion with zero absorption are then feasible .
the interpretation of these results is straight forward via a dressed - state analysis @xcite . when @xmath117 ( see fig .
[ fig-3ab]a , b ) , the probe field at exact resonance with the dressed - state transition @xmath118 . if @xmath119 most population is placed in the dressed - states @xmath120 ( see fig . [ fig-2ab]a ) and , thus , the probe field is absorbed . here
@xmath121 means that @xmath122 while @xmath123 and , respectively , the susceptibility is small though nonzero .
note that for a single - type atomic ensemble one can achieve complete transparency at this point .
increasing further @xmath124 , i.e. @xmath125 , the dressed -state population is transferred completely to @xmath126 and the probe field is amplified .
thus , the second ensemble contributes here to a strong shift of the susceptibility resulting in zero absorption with large dispersive features . in particular , the index of refraction yield with close to vanishing absorption @xmath127 @xcite which for the sample parameters given above takes values larger than @xmath128 [ see fig .
( [ fig-3ab]a ) near @xmath129 .
however , without collective effects , i.e. a noninteracting ensemble of @xmath1 = @xmath130 atoms , the index of refraction would only be close to unity at the point of vanishing absorption
. the index of refraction @xmath131 can be further enhanced by increasing @xmath132 , but then the atoms would be such close to each other that short - range dipole - dipole interactions need be taken into account .
moreover , @xmath133 may be smaller than @xmath134 at frequencies where the absorption vanishes meaning that the weak - probe field can not propagate anymore through the atomic medium ( see fig .
[ fig-3ab ] ) . furthermore ,
once the probe field frequency is adjusted near resonance with the bare - state transition frequencies of atoms @xmath11 or @xmath12 , one can observe steep dispersive features at @xmath135 ( see fig .
[ fig-2ab]b ) . however , the magnitude of the susceptibility , in this case , is the same as for a few - atom sample , except the dispersive slopes .
these results can be employed to induce a rapid phase shift for a weak field travelling through the strongly driven atomic sample .
note that at the points , where the @xmath36 change the sign , the two - level emitters are in a strong collective phase @xcite and , thus , abrupt changes of @xmath133 are due to strong collectivity . also , the collisional damping does not affect considerably the collective steady - state behavior , and its influence can be balanced by increasing the number of atoms .
this means that the collisional damping influences small atomic samples while larger atomic systems are less sensitive with regard to kind of phase damping .
the steady - state dependence of the linear susceptibility @xmath136 ( in units of @xmath137 ) as well as that of the derivation @xmath138 ( in units of @xmath139 ) as a function of @xmath140 .
the solid and dashed lines correspond to the real and imaginary parts of @xmath136 , respectively , while the dotted curve stands for @xmath138 . here
@xmath141 ( a , b ) , and @xmath142 ( c , d ) , while @xmath143 , @xmath144 , @xmath145 , and @xmath146 .
( b , d ) are enlargements of ( a , c).,width=317,height=226 ] we demonstrate further that collections of two - level atoms are suitable for rapidly switching between strongly accelerating or slowing down of a weak probe pulse traversing through driven two - level media .
the light group velocity can be estimated from the following expressions : @xmath147 , where @xmath148 . for @xmath149
, i.e. @xmath150 , @xmath151 @xmath152 may reach values of order of @xmath153 of either sign [ see figs .
( [ fig-3ab]b , d ) where @xmath154 . the refractive index may take values below unity at such moderate atomic densities .
thus , by properly choosing the external parameters one can arrive at rather low subluminal or large superluminal group velocities . in summary we have demonstrated that collective interactions among two - level radiators are suitable to generate highly refractive media with ghz switching times to strongly deviating properties .
further , the group velocity of a weak electromagnetic field pulse probing the laser - driven atomic sample may be abruptly altered depending sensitively on the external atomic and laser parameters . |
associated to any finite graph @xmath2 there is a zeta function @xmath3 , @xmath4 . it is defined as an infinite product but shown ( in various different cases ) by ihara , hashimoto , and bass @xcite to be a polynomial .
indeed the rationality formula for a @xmath5 regular @xmath2 states that : @xmath6 here @xmath7 is the adjacency operator of @xmath2 . in @xcite , an @xmath0-zeta function
is defined for noncompact graphs with symmetries , using the machinery of von neumann algebras . a rationality formula similar to expresses the relationship between the zeta function and the von neumann determinant of a laplace operator .
the results of this paper focus on an especial case .
let @xmath1 be an infinite graph which covers a finite graph @xmath8 .
the @xmath0-zeta function @xmath9 is defined in @xcite only in a small neighborhood of zero .
the first result of this paper is to extend the @xmath0-zeta function to the interior of @xmath10\cup [ \frac 1 q , 1]$ ] .
in the second part of the paper , we consider a tower of finite graphs @xmath11 covered by @xmath1 . put @xmath12 . in theorem
[ thm : conv ] we show that the zeta functions for the @xmath11 , renormalized by taking @xmath13 roots , converge to the @xmath0-zeta function for @xmath1 .
the argument is inspired by , and uses , work of lck @xcite .
in the first section we recall the definitions of the zeta functions of finite and infinite graphs .
one of the main results of this paper is theorem [ extension ] in this section . in the second section of this paper
we prove the convergence theorem and exhibit interesting examples .
theorem [ deit ] generalizes work of deitmar @xcite .
in this section we recall the definition of the zeta function and related material .
we first recall the definition of the zeta function for finite graphs . for a graph @xmath2 , let @xmath14 and @xmath15 denote the sets of vertices and edges , respectively , of @xmath2 .
if each vertex has the same degree then @xmath2 is _
regular_. let @xmath2 be a finite graph .
a closed path in @xmath2 is _ primitive _ if it is not a nontrivial power of another path inside the fundamental group of @xmath2 .
let @xmath16 be the set of free homoptopy classes of primitive closed paths of @xmath2 .
then the zeta function of @xmath2 is @xmath17 where @xmath18 is the minimum length of paths in the class of @xmath19 .
let @xmath7 be the adjacency operator of the graph @xmath2 acting on @xmath20 .
for @xmath21 let @xmath22 where @xmath23 is the degree of the vertex @xmath24 .
put @xmath25 .
the ihara rationality formula says that @xmath3 is a polynomial : @xmath26 the zeta function satisfies the following functional equation : ( * ? ? ?
* corollary 3.10 ) @xmath27 where @xmath28 and @xmath29 . for more details and examples
. the following proposition is well known .
a proof can be found in ( * ? ? ?
* page 59 ) .
the zeros of the zeta function for any finite @xmath5 regular graph lie on the set @xmath30 where @xmath31\cup [ \frac 1 q , 1].\ ] ] let @xmath32 be the interior of @xmath30 . then
: [ omega ] if @xmath33 $ ] then @xmath34 $ ] .
let @xmath35 . for @xmath36 to be real
we must have @xmath37 this implies that either @xmath38 or @xmath39 . if @xmath39 then @xmath40 which is negative or zero if @xmath41 is on or outside of the circle @xmath42 . in the case that @xmath38 put @xmath43 .
if @xmath44 has no real root then @xmath44 is always positive .
otherwise @xmath45 then @xmath46 and @xmath47 are non - negative , and @xmath48 is not zero in @xmath49
. therefore @xmath44 will be positive on @xmath50 .
[ nthroot ] for a @xmath5 regular finite graph x , the polynomial @xmath51 has an analytic @xmath52 root on @xmath32 for all @xmath53 . we know
@xmath54 where @xmath55 varies over eigenvalues of the adjacency operator of the graph . then @xmath56 is an analytic @xmath52 root for @xmath51 .
for @xmath57 a countable discrete group , the _ von neumann algebra _ of @xmath57 is the algebra @xmath58 of bounded @xmath57-equivariant operators from @xmath59 to @xmath59 .
the _ von neumann trace _ of an element @xmath60 is defined by @xmath61 for @xmath62 the unit element . for @xmath63 and a bounded @xmath57-equivariant operator @xmath64 ,
define @xmath65 the trace as defined is independent of the decomposition of @xmath66 .
the von neumann trace extends to bounded @xmath57-equivariant operators on hilbert @xmath58-modules , but they will not be needed .
now let @xmath67 be an infinite graph .
suppose the group @xmath57 acts freely on @xmath1 with finite quotient @xmath68 .
let @xmath16 denote the set of free homotopy classes of primitive closed paths in @xmath1 .
for @xmath69 , @xmath18 is the length of the shortest representative of @xmath19 .
the group @xmath70 is the stabilizer of @xmath19 under the action of @xmath57 .
the _ @xmath0-zeta function _ of @xmath1 is defined in @xcite as the infinite product @xmath71 the adjacency operator @xmath7 and laplace operator @xmath72 are @xmath57-equivariant operators on @xmath73 .
choosing lifts of vertices of @xmath68 yields a decomposition @xmath74 .
then from ( * ? ? ?
* theorem 0.3 ) , @xmath75 in this formula , @xmath76 is the @xmath0-euler characteristic of @xmath1 , which in our setting is simply equal to @xmath77 .
the determinant @xmath78 is defined via formal power series as @xmath79 and converges for small @xmath4 .
more precisely , if @xmath1 is @xmath5 regular then the radius of convergence of @xmath80 is greater than or equal to @xmath81 [ extension ] let @xmath1 be a @xmath5 regular graph
. then @xmath80 has a holomorphic extension to @xmath82 by it is enough to show that @xmath83 has a holomorphic extension on @xmath32 .
let @xmath84 .
here and in the rest of the paper @xmath85 is the principal branch of the logarithm , defined and analytic on @xmath86 $ ] .
fix @xmath87 .
then using proposition [ omega ] , there exists an open set @xmath88 $ ] on which @xmath89 is analytic . since @xmath7 is self - adjoint and @xmath90 , the spectrum @xmath91 $ ] . by the spectral theorem for self - adjoint operators we can write : @xmath92 now @xmath93 is well defined , and @xmath94 is a holomorphic function of @xmath41 on @xmath32 .
now for small @xmath41 , @xmath95 using the functional equation it is possible to extend the @xmath96-zeta function to the exterior of @xmath30 . defining the zeta function on @xmath30 itself presents some problems . in examples ,
the function on the interior of @xmath30 and the exterior of @xmath30 do not match continuously on @xmath30 .
however the absolute value of the resulting function is the fuglede - kadison determinant , and may extend continuously to @xmath30 .
in this section we prove that the zeta functions for a tower of finite graphs , suitably renormalized , converge to the @xmath0-zeta function for an infinite covering graph .
the argument uses an idea from @xcite .
[ thm : conv ] let @xmath1 be a @xmath5 regular graph .
suppose the group @xmath57 acts freely on @xmath1 , and that @xmath8 is a finite graph .
suppose @xmath97 is a tower of finite index normal subgroups and @xmath98 .
let @xmath99=n_i$ ] and @xmath100 .
then for @xmath101 , we have @xmath102 the convergence is uniform on compact subsets of @xmath32 . from the remark
we need to show that @xmath105 let @xmath106 .
let @xmath107 , where @xmath108 is the spectral decomposition of @xmath7 acting on @xmath109 we now set @xmath110 then from ( * ? ? ?
* theorem 2.3.1 ) , for all @xmath111 $ ] , @xmath112 we know @xmath113 by @xmath114 if @xmath115 is compact then @xmath116 is bounded uniformly for @xmath117 and @xmath55 in an open interval containing @xmath118.$ ] now integration by parts shows that indeed @xmath119 as @xmath120 .
let @xmath1 be the line , with vertices @xmath121 and edges connecting @xmath122 to @xmath123 .
the group @xmath124 acts on @xmath1 with quotient having one vertex and one edge .
let @xmath125 , so @xmath126 is an @xmath122-cycle .
as @xmath127 has only two primitive loops and @xmath1 has none , we have @xmath128 and @xmath129 .
the graphs are 2-regular , so @xmath32 is the ( open ) unit disk . for @xmath101 , @xmath130 notice that the functional equation gives @xmath131 outside of the unit disk .
let @xmath1 be the @xmath5 regular tree and @xmath132 finite .
since @xmath1 has no closed loops , the @xmath0-zeta function of @xmath1 is the constant function 1 .
so if @xmath126 is a tower of finite graphs covering @xmath68 , @xmath133 } = 1\ ] ] for @xmath101 .
this result is contained in @xcite for @xmath134 small .
k. hashimoto .
zeta functions of finite graphs and representations of @xmath140-adic groups . in _ automorphic forms and geometry of arithmetic varieties _ , volume 15 of _ adv .
studies in pure mathematics_. academic press , 1989 . | the @xmath0-zeta function of an infinite graph @xmath1 ( defined previously in a ball around zero ) has an analytic extension . for a tower of finite graphs covered by @xmath1 ,
the normalized zeta functions of the finite graphs converge to the @xmath0-zeta function of @xmath1 . |
dozen years after its unexpected and somewhat serendipitous discoverty , the accelerated expansion of the universe is taken for granted due to the flood of data from different astrophysical probes confirming it ( see , e.g. , @xcite for a not updated yet exhaustive review ) .
although the spatially flat concordance @xmath4cdm model , made out of a cosmological constant accounting for @xmath5 of the energy budget and responsible of the cosmic speed up , is in full agreement with observations ( as summarized , e.g. , in @xcite , where the latest datasets are considered ) , it is far from free of any conceptual and theoretical problems @xcite .
this motivated the search for alternative models generally referred to as dark energy ( de ) ones and characterized by the presence of a scalar field with a negative equation of state ( eos ) denoted as @xmath6 , with @xmath7 the pressure and energy density of this leading component . while a huge number of papers has addressed the problem of what de is with proposal running from self interacting fields to modified gravity theories ( see , e.g. , @xcite for a recent textbook ) , observations are mainly aimed at constraining the de eos . under the simple yet efficient cpl @xcite parameterization
, it is considered @xmath8 , with @xmath9 the scale factor and @xmath10 the redshift , so that the holy grail of observational cosmology has nowadays become to narrow down as much as possible the range for the @xmath11 parameters . as a consequence , the efficacy of an observational probe
is presently quantified in terms of the figure of merit ( fom ) defined as the inverse of the area delimiting the @xmath12 confidence range in the @xmath11 space @xcite .
gravitational lensing has soon been pointed at as one of the most promising tools to investigate the nature of the mysterious de . in its journy from the source to the observer
, the light is deflected by both a dominant mass concentration ( a galaxy or a cluster acting as lens ) along the line of sight and the cumulative effect of the large scale structure . as a consequence , gravitational lensing probe both the background expansion and the growth of structures .
cosmic shear tomography , that is to say the power spectrum of shear as measured from galaxies separated in different redshift bins @xcite , has emerged as one of the most promising tools to constrain the de eos , so motivating the interest in ongoing ( des , kids ) and future ( euclid , lsst ) dedicated surveys . while shear tomography is based on the weak regime of gravitational lensing ( where lensed galaxies are subject to tiny modifications detectable only on statistical grounds ) , also statistics based on strong lensing features
is not less important . in this regime , the light emitted by the source passes close to a massive lens causing the formation of multiple images or spectacular einstein rings or radial and/or tangential arcs .
the distribution of angular separation between double images @xcite , arc statistics @xcite and einstein rings properties @xcite have been used as tools to constrain cosmological parameters and the astrophysical properties of the lens population .
galaxy scale strong lenses have been also proposed as a tool to probe the background expansion .
the basic idea is usually to infer the distance ratio @xmath13 between the angular diameter distance from the lens to the source and from the observer to the source , and then use a sufficiently large sample to trace how this quantity evolves with @xmath10 . to this end ,
one typically relies on measurements of both the einstein radius @xmath14 and the lens velocity dispersion @xmath15 . under the assumption of isothermal sphere
, one gets @xmath16 , so that a measurement of @xmath17 immediately gives the distance ratio @xmath13 provided one sets @xmath18 , being @xmath19 a nuisance parameter to marginalize over @xcite .
although quite simple , such a method however rests on some assumptions which , even if reasonably motivated , should nonetheless be verified .
first , it is assumed that all the lenses can be modelled as singular isothermal sphere .
second , the nuisance parameter @xmath19 is taken as a sort of universal constant , while its actual value can change from one system to another depending on both the lens properties ( redshift , mass , size ) and how @xmath15 is measured ( weighted within a circular aperture of fixed physical radius or through longslit spectroscopy ) . here
we follow a similar approach , but weakening the strong assumptions of the standard approach in order to avoid biasing cosmological parameters because of incorrect lens modelling .
the plan of the paper is as follows . in sect.2 , a step - by - step derivation of the method
is presented underlying which assumptions are made and how we deal with them .
sect.3 is devoted to validate the proposed technique showing that its basic assumptions are well motivated .
we then present , in sect.4 , a fisher matrix analysis to forecast the errors on cosmological parameters for different configurations of future datasets .
a summary of the results and a discussion of future perspectives are given in the concluding sect.5 .
galaxy scale strong lenses are typically used to constrain the dark matter halo and its interplay with the properties of the stellar component . to this end ,
two observational probes are used , namely the projected mass @xmath20 within the einstein radius @xmath14 and the central velocity dispersion @xmath15 . assuming the galaxy may be modelled as the sum of a stellar part modelled with a sersic ( 1968 ) profile and a dark matter halo , we get @xmath21\ , , \label{eq : mprojth}\ ] ] with @xmath22 the total stellar and virial masses , respectively , and @xmath23 is a dimensionless function depending on the density profile . for the stellar component this reads @xmath24}{\gamma(2n)}\ , , \label{eq : defmustar}\ ] ] with @xmath25 the slope of the sersic model , @xmath26 a constant set so that half of the total luminosity is contained within the effective radius @xmath27 @xcite .
we leave for the moment the @xmath28 function undetermined in order to be as general as possible . provided that the lens and source redshifts @xmath29
are known , one can get an estimate of the projected mass by simply measuring the einstein radius and then using @xmath30 where we have used the definition of critical surface density @xmath31 : @xmath32 being @xmath33 the speed of light and @xmath34 the angular diameter distances to the lens , the source and between lens and source .
note that , in eq.([eq : mprojobs ] ) , @xmath35 is the einstein radius in arcsec since this is the quantity directly measured from the data in a model independent way .
the second quantity of interest is the aperture velocity dispersion , i.e. the velocity dispersion luminosity weighted within a circular aperture of radius @xmath36 .
approximating the deprojected sersic profile with the @xcite model , it turns out that @xmath37 } \\ & \times & \frac{{\cal{i}}_{\star}(r_{ap } ; n ) + ( m_{vir}/m_{\star } ) { \cal{i}}_{dm}(r_{ap } ; n , r_{eff } ; { \bf p}_{dm } ) } { \gamma(2n ) - \gamma[2n , b_n ( r_{ap}/r_{eff})^{1/n } ] } \nonumber \ , \label{eq : sigmaapth}\end{aligned}\ ] ] where @xmath38 is given in @xcite , @xmath39 collectively denotes the halo model parameters , and we refer to @xcite and @xcite for the details of the dimensionless @xmath40 functions . for dimensional reasons it is convenient to define an aperture mass as @xmath41 } \\ & \times & \frac{{\cal{i}}_{\star}(r_{ap } ; n ) + ( m_{vir}/m_{\star } ) { \cal{i}}_{dm}(r_{ap } ; n , r_{eff } ; { \bf p}_{dm } ) } { \gamma(2n ) - \gamma[2n , b_n ( r_{ap}/r_{eff})^{1/n } ] } \nonumber \ , \label{eq : mapth}\end{aligned}\ ] ] which can be straightforwardly estimated from measurable quantities as @xmath42 with @xmath43 the effective radius in arcsec obtained from the fit to the surface brightness profile .
it is now only a matter of algebra to first solve @xmath44 with respect to @xmath45 and then insert the result into @xmath46 to finally get @xmath47 which represents the basic relation of our method relating the theoretical distance ratio @xmath48 on the left hand side with the @xmath49 quantities on the right hand side defined as follows @xmath50}{b_n^{n(3 - p_n ) } } \nonumber \\ & \times & \left \ { 1 - \frac{\gamma[2n , b_n ( r_{ap}/r_{eff})^{1/n}]}{\gamma(2n ) } \right \ } \nonumber \\ & \times & \left \{1 - \frac{\gamma[n(3 - p_n ) , b_n ( r_{ap}/r_{eff})^{1/n}]}{\gamma[n(3 - p_n ) ] } \right \ } \ , \label{eq : defdobs}\end{aligned}\ ] ] @xmath51 } { { \cal{i}}_{\star}(r_{ap } ; n ) + ( m_{vir}/m_{\star } ) { \cal{i}}_{dm}(r_{ap } ; n , r_{eff } ; { \bf p}_{dm } ) } \ . \label{eq : defcalfe}\ ] ] eq.([eq : testeq ] ) is formally similar to the one used in previous similar methods expressing the distance ratio in terms of a quantity depending only on measurable quantities ( given by @xmath52 ) and an unknown nuisance parameter ( referred to as @xmath53 here ) .
however , we improve on the classical approach in two ways . first , we have not assumed the galaxy to be a one component system , but rather modelled it as it actually is , i.e. the sum of a stellar part and a dark matter halo .
second , we have now an analytical expression for the nuisance parameter , so that we can investigate whether it is the same for all lenses or rather it is a function of both the lens properties and the halo model .
to infer the distance ratio from strong lensing data , one generally assumes that the @xmath19 parameter is a constant independent on the lens and marginalizes over it .
this is actually a zero order approximation related to the properties of the assumed singular isothermal sphere model used to describe the total mass density . on the contrary , with our @xmath53 parameter in eq.([eq : defcalfe ] ) we account for the two components nature of the lens and explicitily specify the halo model .
a careful analysis of the different terms entering its definition shows that @xmath53 depends on ( i ) the mass ratio @xmath54 , ( ii ) the length ratio @xmath55 ( with @xmath56 the radius where the dm logarithmic slope equals -2 ) , and ( iii ) the halo concentration provided that the logarithmic slope is not constantly equal to -2 , which is the case for the singular isothermal sphere only . ]
actually , the halo model can be more complex , so that @xmath58 are not enough to fully assign it , and more parameters are needed , which makes @xmath53 depend on them too .
moreover , since the mass and length ratios may both change with the redshift as it is suggested by , e.g. , the redshift dependence of the @xmath59-@xmath60 relation as well as of the @xmath27-@xmath45 scaling law , it turns out that @xmath53 is redshift dependent .
these considerations strongly point against the standard assumption to consider the nuisance parameter as a constant , which implies that one should fit for @xmath53 separately for each lens , hence making the method unusable .
however , a possible way out could be found if @xmath53 was constant within a certain scatter once the lenses were binned according to some suitably chosen properties . to investigate whether this is possible or not
, we proceed as schematically pointed out below . 1 .
we set the lens redshift @xmath61 sampling from a uniform distribution over the range @xmath62 and then set @xmath63 with @xmath64 randomly extracted from the range @xmath65 .
we use the redshift dependent galactic stellar mass function used in @xcite to set the lens stellar mass @xmath45 and sample the effective radius @xmath27 from a gaussian distribution centred on @xmath66 and a @xmath67 scatter .
the estimation of @xmath68 is obtained following @xcite , while the scaling with redshift has been chosen in accordance with @xcite considering the value of @xmath69 for massive galaxies ( @xmath70 ) .
the halo virial mass is set solving the @xmath71 vs @xmath60 relation of @xcite with respect to @xmath60 , so that we can finally fix the concentration sampling from the @xmath59-@xmath60 relation of @xcite , taking also into account its scatter .
+ 4 . adopting an nfw @xcite model for the dm halo
, we can now compute both the einstein radius @xmath14 and the central velocity dispersion @xmath72 .
we then retain the simulated lens only if its input and output quantities are reasonable . to check whether this is the case
, we compare them to the similar quantities for the slacs @xcite sample of 85 intermediate redshift lenses with well measured values of @xmath73 . using the above recipe
, we can generate a sample of @xmath74 lenses , with a distribution in @xmath73 similar to those of the real slacs , and confidently use it to investigate how the nuisance parameter , @xmath53 , scales with the lens properties .
moreover , since we are interested more in the distance ratio estimate than in @xmath53 , we wonder what could be the bias on @xmath75 caused by using the mean value @xmath76 , averaged over a subsample of 1000 lenses selected according to a given criterion , instead of @xmath53 . to this end
, we define the distance ratio estimator @xmath77 and consider the quantity @xmath78 as a measure of the relative error due to the distance ratio estimate . since @xmath79 is the average value over the lenses in a given bin ( so that we can trace how @xmath80 evolves with the bin centre ) , the question is now which parameter to use as binning quantity . as a first case
, we consider the most obvious choice and divide the simulated sample in 50 redshift bins .
it turns out that , while @xmath79 increases with @xmath61 , its relative scatter only has a negligible variation , and that the root mean square value is @xmath81 independently of @xmath61 .
moreover , we have verified that , for lenses in the same bin , @xmath80 has a negligible correlation with the source redshift , the lens size and mass , and the halo model parameters .
we repeat this test , binning with respect to the stellar mass , central velocity dispersion , and einstein radius .
we find that we are unable to get satisfactory results . indeed , binning with respect to @xmath82 increases @xmath83 and introduces a significant increase with the stellar mass for @xmath84 , with values going as large as @xmath85 .
similarly large values are obtained when binning with respect to @xmath15 although a definite trend is not found , while binning in @xmath14 introduces an increasing trend for @xmath86 followed by a non monotonic variation leading to @xmath87 ( but the low statistics at large @xmath14 values makes these numbers not fully reliable ) .
we therefore conclude that the estimator ( [ eq : hatd ] ) is reliable , once the lenses are divided in redshift bins . in this case , one can then rely on @xmath88 assuming that the nuisance parameter is the same for all the lenses in the same bin within a percentage scatter of @xmath89 . to further validate this conclusion ,
we have repeated the full analysis changing the assumptions in the simulation pipeline .
we have indeeed considered a different @xmath59-@xmath60 relation either by adopting the one of @xcite or by changing the halo model from nfw to einasto ( 1965 , 1969 ) profile .
although the @xmath53 values are different , the results on @xmath83 are qualitatively and quantitatively similar .
in particular , we confirm the above conclusion that binning in lens redshift guarantees that @xmath81 independently of the bin centre and the stellar and dm halo properties of the lenses .
in order to investigate whether and under which conditions the method proposed works in constraining the cosmological parameters , we perform a fisher matrix analysis .
the elements of the fisher matrix are given by @xmath90 being @xmath91 the i - th parameter and @xmath92 the fiducial values .
to define the likelihood function @xmath93 we split the sample in @xmath94 bins and let @xmath95 be the number of lenses in the @xmath96-th bin .
we can compute the likelihood for lenses in this bin as @xmath97 ^ 2 } \label{eq : deflikek}\end{aligned}\ ] ] where @xmath98 is the set of cosmological parameters to be constrained , while @xmath99 is the average @xmath53 value for the @xmath100 bin which we marginalize over . in order to compute the likelihood ( and its derivatives ) we need to assign the quantities entering eq.([eq : deflikek ] ) .
the @xmath101 may be easily computed for each lens using eq.([eq : defdobs ] ) ; the errors @xmath102 deserve , on the other hand , some more words .
first of all , @xmath103 is the error on @xmath104 . since we are assuming that @xmath99 is constant , we can then simply write @xmath105 , where a naive propagation of errors is sufficient to get @xmath106 here @xmath107 are the relative uncertainties on @xmath108 , respectively
. we will make the simplifying assumption that these quantities are the same for all the lenses in a given bin and explore different values for them ( see later ) .
the second term @xmath109 is the systematic error induced by replacing the actual @xmath53 with its average value and can be set as @xmath110 .
based on the simulations described in the previous section , we will set @xmath111 , while @xmath99 is computed by binning the lenses according to the above chosen strategy .
assuming that the lens redshift has been measured with a precision much smaller than the bin width , the total likelihood is simply : @xmath112 where we collectively denote with @xmath113 the set of @xmath99 values . as a consequence
, the total fisher matrix will be the sum of the fisher matrices for each bin .
it is convenient to first compute the fisher matrix for the @xmath96-th bin , then marginalize over @xmath99 , and finally sum the marginalized matrices to get @xmath114 , i.e. , the total fisher matrix for the cosmological parameters only . in order to compute the fisher matrix
, some preliminary quantities must be set .
first , we choose the fiducial cosmological model .
we consider three different spatially flat models setting @xmath115 which we will refer to as @xmath4cdm , quiessence , thawing , respectively .
note that we have used the same matter density parameter for all three models , setting its value to the recent planck estimate although one should change it according to the de eos parameters in order to get the best match to the data .
however , we are here only interested in exploring realistic models so that we do not worry much about the exact value of the parameters .
we nevertheless want to stress that these cases span the range of quintessence models still allowed by the data @xcite .
these parameters assign the dimensionless hubble parameter @xmath116 , reading @xmath117 which enters the distance ratio given by @xmath118 with @xmath119 the comoving distance .
note that eq.([eq : distratio ] ) only holds for spatially flat models .
we also stress that the hubble constant @xmath120 drops out from the ratio of distances thus reducing the number of cosmological parameters . a key role in determining the accuracy of the constraints
is played by the errors on the observable quantities .
these enters eq.([eq : deflikek ] ) through the statistical term @xmath121 and the systematic one @xmath109 . as already said above , we set @xmath110 , since this is the scatter in the distance ratio introduced by the approximation done when using the estimator @xmath88 instead of the correct one in eq.([eq : testeq ] ) . on the other hand ,
the statistical uncertainty @xmath121 depends on the relative errors on @xmath108 .
we will set @xmath122 as typically found fitting hst - like data on lens surface brightness profile , while we try three different configurations for the uncertainties on the einstein radius and central velocity dispersion , namely @xmath123 as a final ingredient , we need to set the total number of lenses and the binning .
we consider two cases .
first , we take 1000 lenses splitted in 20 equally spaced bins over the redshift range @xmath62 , while secondly we take 10000 lenses splitted in 50 bins over the same redshift range as improved sample .
.marginalized constraints on @xmath124 for @xmath4cdm , quiessence and tahwing models ( top , centre and bottom part of the table , respectively ) for different sample and error configurations .
the labels @xmath125 and @xmath126 refer to the sample containing 1000 and 10000 lenses , respectively .
we remind the reader that we set @xmath122 for all the cases . [
cols="^,^,^,^,^,^",options="header " , ] [ tab : fishmatrescst ] table[tab : fishmatrescst ] and [ tab : fishmatresvar ] summarizes the results of the fisher matrix forecast for different models , sample and error configurations .
first , we consider the simplifying assumption that @xmath127 is known @xcite . ] . while for @xmath4cdm and quiessence models this means that , in a hypothetical fit , we force @xmath128 , in the case of thawing model , one has to set @xmath129 in order to get the correct behaviour @xcite .
the numbers in table[tab : fishmatrescst ] convincingly show that the distance ratio method is quite effective at constraining the cosmological parameters @xmath124 if a prior on @xmath127 is set .
no matter which model is taken as fiducial , we find that 1000 lenses are sufficinet to get @xmath130 and @xmath131 .
increasing the sample by an order of magnitude roughly halves these numbers pointing at a possible saturation of the method accuracy . for fixed model and lens sample ,
it turns out that reducing the error on @xmath15 is a better strategy to improve the constraints since it is more effective at narrow down the error on @xmath132 .
the constraints are basically the same for the three models , although those for the thawing case are consistently smaller .
this can be qualitatively explained noting that , in this case , @xmath132 has a larger impact on the distance ratio thanks to the presence of the exponential term in the dimensionless hubble parameter ( which drops out when @xmath128 is used ) .
although encouraging , these results are based on the strong prior that we know how to set @xmath127 . in a realistic application , however , one only assumes that the de eos is given by the cpl approximation and fits for the three parameters @xmath133 .
the constraints we get are summarized in table[tab : fishmatresvar ] and show a significant degradation with respect to the case with @xmath127 fixed .
while the trends with @xmath134 are qualitatively the same , the number of lenses becomes now of primary importance .
indeed , 10000 rather than 1000 lenses are now needed to get @xmath130 and @xmath135 , while only weak constraints can be put on @xmath127 .
this result can be easily understood looking at the pivot redshift @xmath136 defined as the value of @xmath10 where the errors are uncorrelated .
not surprisingly , we find that @xmath137 no matter which error configuration , lens sample or fiducial model is used .
this is an obvious consequence of the lens redshift distribution which indeed has a median value close to the pivot redshift . as a result ,
the de eos is best constrained at @xmath136 with the error on @xmath138 turning out to becdm , but they are quite similar for other models . ]
@xmath139 from 1000 lenses in the three error configurations and @xmath140 from 10000 lenses .
these results clearly show that the distance ratio method is better suited at probing the low redshift behaviour of the de eos so that its redshift evolution is hardly probed thus explaining the weak constraints on @xmath127 .
it is interesting to look at the fom of the method .
we get @xmath141 when using 1000 lenses , while it is @xmath142 if 10000 lenses are used .
it is evident that interesting values can only be obtained reducing as much as possible the error on @xmath15 and increasing the lens sample .
we , however , note that the orientation of the @xmath11 ellipses ( see fig,[fig : contplot ] ) , is quite different from other classical methods such as sneia and cmbr thanks to the different pivot redshift .
it is therefore likely that a combination of the distance ratio method with these other probes helps breaking degeneracies thus boosting the total fom .
[ tab : fishmatresvar ]
einstein rings have always attracted a lot of attention not only for their spectacular beauty , but also as a probe of the dark matter content of lens galaxies .
the measurement of both the einstein radius and central velocity dispersion allows to strengthen the constraints on the halo density profile , but can also be used as a way to estimate the ratio between the distance to the source and that between lens and source and hence probe the cosmic expansion .
motivated by the previous literature results , we have proposed a novel approach which , although based on the same idea , ameliorates the estimate of the distance ratio . on one hand ,
we have adopted a realistic two components model for the lens galaxy taking care of both the stellar and dm contributions .
this led us to a general formula which permits to estimate the distance ratio @xmath143 as a function of observable quantities and a nuisance parameter @xmath53 depending on the details of the halo model and the lens parameters .
the analytical formula we derive for @xmath53 has allowed us to show that its value is approximately constant for lenses belonging to the same redshift bin , so that we finally get an approximate estimator which helps us to define a likelihood function for the estimate of cosmological parameters .
it is worth stressing that the method proposed does not depend on the particular halo model adopted or on the details of the @xmath71-@xmath45 and @xmath59-@xmath60 relations adopted .
this is a further improvement with respect to the standard approach which , on the constrary , postulates that all lenses may be described by the singular isothermal sphere
. a well founded and movitated method to constrain cosmological parameters should be useless if the numbers of tracers and/or the requirements on the uncertainties on the observable quantities are too demanding . to investigate this ,
we have carried on a fisher matrix analysis changing both the number of lenses and the relative errors on @xmath144 .
it turns out that a survey measuring @xmath145 with @xmath146 accuracy for 10000 lenses can get @xmath147 , while @xmath148 constraints on @xmath124 can be obtained with 1000 lenses only if @xmath127 is fixed .
it is therefore worth wondering whether these requirements are realistic . to this end
, we note that present day lenses already achieved @xmath149 and @xmath150 , so that it is not unrealistic to guess that higher quality images ( as can be obtained from future large telescopes and satellite missions ) and improved spectrograph can easily achieve our requirements .
more demanding is the constraint on the total number of lenses . as an example , however , one can note that the euclid mission @xcite is expected to observe up to @xmath151 strong lenses @xcite , so that asking that both @xmath14 and @xmath15 are measured for less than @xmath67 of them is not a too demanding requirement . although the present day lens samples are far from the numbers explored in our fisher matrix analysis , it could be nevertheless interesting to apply the proposed method to real strong lensing data . @xcite
have recently assembled a catalog of 118 strong lensing systems and used it to constrain the dark energy equation of state under the usual assumption of isothermal sphere model .
we plan to repeat their analysis using our improved approach in order to investigate to which extent the constraints on the cosmological parameters depend on the assumed lens model .
actually , given the small statistics , it is expected that the large errors could prevent a conclusive answer , but such a study can provide a comparison benchmark for an analysis based on simulated samples .
it is worth noting that the results are always somewhat dependent on the lens redshift distribution . here
, we have simulated lens samples in such a way that they grossly reproduce the main features of the slacs dataset since this is the most used catalog at the moment .
however , future datasets can have a different redshift distribution , possibly changing the method fom .
although a more detailed analysis will be carried on in a future work , we can qualitatively anticipate that the pivot redshift will likely be close to the median survey redshift .
it is therefore possible that combining different samples with different median redshifts could help better tracing the de eos over a larger redshift range thus increasing the fom .
note that we are here implicitly assuming that the median redshift of the full lens sample is the same as the median redshift of the subsample with measured values of @xmath152 , which can also not be the case because of selection effects . on the contrary , since not all the lenses contribute to the fom in the same way , it is possible that an optimal strategy can be worked out maximizing the scientific return of the companion spectroscopic survey needed to measure the lenses velocity dispersions .
to this end , one can follow the same methodology presented , for time delay distances , in @xcite . as a final remark ,
we stress that , even if suboptimal conditions in terms of number of lenses and/or accuracy in the measurement of @xmath145 are achieved , the distance ratio method we have proposed can be efficiently combined with other probes helping to break degeneracies and hence boosting the total fom . should the present results be confirmed by a more detailed analysis of the lens samples properties , new light on dark energy will be shed by the strong regime of gravitational lensing .
vfc is funded by italian space agency ( asi ) through contract euclid - ic ( i/031/10/0 ) and acknowledges financial contribution from the agreement asi / inaf / i/023/12/0 . | strong lensing provides popular techniques to investigate the mass distribution of intermediate redshift galaxies , testing galaxy evolution and formation scenarios .
it especially probes the background cosmic expansion , hence constraining cosmological parameters .
the measurement of einstein radii and central velocity dispersions indeed allows to trace the ratio @xmath0 between the distance @xmath1 from the observer to the source and the distance @xmath2 from the lens to the source .
we present an improved method to explicitly include the two - component structure in the galaxy lens modeling , in order to analyze the role played by the redshift and the model dependence on a nuisance parameter,@xmath3 , which is usually marginalized in the cosmological applications .
we show how to deal with these problems and carry on a fisher matrix analysis to infer the accuracy on cosmological parameters achieved by this method .
gravitational lensing
cosmology : distances |
graphene is a promising material for studying quantum hall effects with gate - tunable filling factors on account of its capability for controlling charge density via field effects@xcite .
studies involving conductance measurements through graphene under a homogeneous magnetic field report non - integer conductance plateaus for gate - tunable bipolar junctions@xcite .
the topological nature of the quantum hall system has been clearly understood via the presence of robust edge states which are topologically protected@xcite .
the bulk - boundary correspondence gives us an intuitive way of understanding the properties of these topological edge states : the number of conducting channels is characterized by the topological invariant of the quantum hall insulator@xcite .
it has been well known that the topological invariant ( or so - called chern number ) of a quantum hall insulator is given by the filling factor in the integer quantum hall effect@xcite .
the observation of non - integer conductance plateaus in bipolar graphene quantum hall systems has been interpreted by the equilibration of topological states at the interface , and theoretical efforts have supported experimental findings by considering disorders on the edges and interface@xcite . despite the agreement between theoretical supports and experimental measurements for small filling factor cases , there are still open questions for relatively larger filling factors namely , the incompleteness of the equilibration and its origin .
recently , it has been theoretically revealed that the equilibration of the topological states at a chern insulator junction can deteriorate if the junction is not perfectly fabricated , thereby limiting the interplay between higher - order landau level ( ll ) states@xcite .
junction conductance via interface equilibration has also been reported for p - n - p junctions in quantum hall graphene systems@xcite , with the consideration that there can be reflections at the bipolar junction . despite efforts to understand the interplay behaviour of the topological interface states using multi - terminal conductance@xcite , verification of the predicted reflections at the junction interfaces
is still desired and can be obtained by delving into exactly how the states behave at a chern insulator junction . in this study ,
a four - terminal geometry is considered to allow the examination of the longitudinal and quantum hall conductance via different leads , individually . as the states are topologically protected , they are robust against disorders which results in a lack of backscattering .
thus , it is straightforward to predict that the states are split into two directions along the edges . even in two - terminal conductance studies , the splitting of the topological interface states
is also expected but is indirectly reflected in the resulting conductance plateaus across junctions .
the four - terminal geometry considered in this work allows us to independently detect both left- and right - split modes from the topological interface states via two individual leads ; additionally , the typical edge states are also measured via the longitudinal conductance ( see fig .
, the sum of the two split conductances must be equal to the typical quantum hall conductance because of flux conservation .
we also take into account random roughness effects and model a more realistic profile of the junction .
overall , the four - terminal conductance with flux conservation of the topological interface states provides an intuitive way of understanding how the interplay between the states occurs at the junction interface .
at @xmath0 t. the dashed line represents the given fermi energy @xmath1 and the dot marks the intersection of the fermi energy and the eigenenergies , indicating the existence of interface states .
( d ) local density of states of the same system as ( c ) .
the light and dark colours represent low and high densities , respectively.,width=321 ] in terms of topology , the bipolar junction of graphene in the quantum hall regime can be regarded as a chern insulator junction , which is feasible to fabricate by applying a homogeneous magnetic field to graphene where a bipolar junction is introduced via gated structures@xcite(see fig .
[ fg : repfigs ] ) . with an analytical approach ,
the effective dirac hamiltonian for graphene under a homogeneous magnetic field reads @xmath2 where @xmath3 ms@xmath4 is the fermi velocity of graphene , @xmath5 are the pauli matrices , @xmath6 , and the electrostatic potential is given by @xmath7 which becomes an abrupt potential step for @xmath8 . the dirac equation with the above effective hamiltonian , @xmath9 with @xmath10
, consists of two sublattice - coupled equations and becomes analytically solvable by decoupling them , resulting in the following schrdinger - like second - order differential equation : @xmath11^{2}\right\}\psi_{a , b}=0 , \label{eq : effham}\end{aligned}\ ] ] where @xmath12 for different sublattices , and @xmath13 is the @xmath14-component momentum which acts as a good quantum number since @xmath15=0 $ ] .
the above equation is dimensionless upon @xmath16 and @xmath17 , and we choose the landau gauge , i.e. , @xmath18 , which leads to @xmath19 .
( [ eq : effham ] ) can be regarded as if dirac fermions experience the effective potential given by @xmath20 , and the solutions for @xmath21 and @xmath22 are obtained as parabolic cylinder functions in the low - energy limit@xcite : @xmath23 where @xmath24 for each region with the potential @xmath25 , @xmath26 , and @xmath27 .
note that the normalization factor is found to be @xmath28^{1/4}$ ] .
alternatively , the tight - binding approach also accounts for the present system , leading to the following hamiltonian : @xmath29 where @xmath30 is the on - site energy including the potential step ( eq . ( [ eq : bipolarcontinuum ] ) ) , and @xmath31 and @xmath32 are the creation and annihilation operators on the @xmath33-th site . in the presence of a magnetic field ,
the hopping term is defined by @xmath34 where @xmath35 = 3.0 ev is the hopping energy , and @xmath36 is the flux quantum . here , we choose the same gauge considered in the analytical approach .
note that the system is centered at @xmath37 . according to the defined potential profile ,
the system is divided into two regions where the lowest lls are located at @xmath38 and @xmath39 .
the ballistic conductance of the four - terminal graphene hall bar is calculated in the linear response regime , exploiting the landauer - bttiker approach , as @xmath40 where @xmath41 is the scattering matrix from lead @xmath42 to @xmath43 . with the tight - binding hamiltonian
, we can also take local density of states ( ldos ) at a given energy : @xmath44 where @xmath45 is the eigenfunction of the dirac hamiltonian , and @xmath46 specifies @xmath47 orbital eigenstates of graphene with energy @xmath48 .
in addition , the probability density of the scattering region for the incoming wavefunction through a given lead is calculated by @xmath49 to provide useful information on how the current actually flows .
both analytic and tight - binding approaches to understanding the electronic and transport properties of the system are in agreement with each other .
in particular , formation of the interface states at the bipolar junction is demonstrated by the eigenenergy bands , which are obtained by using wavefunction continuity conditions at the interface , @xmath50 ( see fig .
[ fg : repfigs](c ) ) . for given potentials @xmath51
, one can see that a conducting channel must exist between electron and hole - doped regions at @xmath1 , connecting the same - index lls .
for the same potential , the existence of the interface channels is also ensured by the ldos at @xmath1 from the tight - binding approach using the kwant code@xcite .
it is worth mentioning that the occurrence of the metallic channels along the interface of the bipolar junction is due to the bulk - boundary correspondence ; therefore , the interface states are topologically protected .
the number of topological interface states is in keeping with the chern number configuration of the bipolar junction in the quantum hall graphene system .
in fact , the chern number of the quantum hall insulator is derived from the tknn formalism@xcite , and the chern number of each region turns out to be the filling factor .
note that , here , we define the odd - number filling factor of the graphene quantum hall regime in units of @xmath52 , distinct from the even - number filling factor in units of @xmath53 . since the topological interface states are doubly degenerate because of valley symmetry , the single channel at the interface in both figs .
[ fg : repfigs]c and d actually contains two topological states , and the number of the topological interface states equals the chern number difference for @xmath54 . as reported in previous studies@xcite , in bipolar quantum hall graphene , the topological interface states exist only in case of p - n junctions on the basis of the equilibration concept .
for a two - terminal geometry , conductance measurements through the bipolar junctions are attributed to either the mode - mixing between parallel - propagating states at the p - n junction interface or edge modes along the system boundary in case of n - n ( p - p ) junctions , according to the polarity of the bipolar junctions .
since the system is finite , the conducting channel along the interface states ends up split into opposite directions , resulting in reflection modes that have been indirectly measured@xcite . on the other hand , using four - terminal geometry , here we are able to measure the splitting of the conductance through the topological interface states with two individual transverse conductances @xmath55 and @xmath56 which are taken between leads 1 and 3 or 4 , as displayed in fig .
[ fg : repfigs](b ) .
in addition , we also take into account the longitudinal conductance @xmath57 to detect the edge state contribution to the conductance in case of n - n ( p - p ) junctions . , depending on the fermi energy .
the dashed lines display quantized conductance values .
( b ) , ( c ) , and ( d ) probability densities of the propagating modes , coming in from lead 1 , at fermi energies @xmath1 , 0.15 , and -0.15 ev , respectively .
the four leads are identical to those depicted in fig .
[ fg : repfigs](b ) .
the applied magnetic field strength is approximately 30 t , being consistent with fig .
[ fg : repfigs](c).,width=321 ] figure [ fg : cleancond](a ) shows the calculated conductances for the given bipolar junction @xmath51 as functions of fermi energy . in this case
, the bipolar junction can be regarded as a chern insulator junction where a topological interface is formed between @xmath58 and @xmath59 regions .
when @xmath60 , the topological states are formed at the junction interface , resulting in @xmath55 and @xmath56 along the interface which are split into the individual leads .
one can clearly see that the sum of the two individual transverse conductances becomes @xmath52 , which is equal to the quantum hall conductance plateau from the lowest - ll states@xcite . besides
, it is seen that @xmath57 vanishes for @xmath60 , because the incoming mode wholly propagates along the interface , as shown in fig .
[ fg : cleancond](b ) . for @xmath61 , the longitudinal conductance @xmath57 from lead 1 to 2
is also expected to be zero due to the propagation of the edge states along the system boundary in the opposite direction , i.e. , from lead 2 to 1 . by
the same token , @xmath55 and @xmath56 also vanish for @xmath61 with the reversal of the propagating direction of the edge states in the whole system , as displayed in fig .
[ fg : cleancond](c ) .
interestingly , for @xmath62 , while @xmath55 exhibits typical quantum hall conductance plateaus as @xmath63 varies , the plateaus are smaller by a factor of @xmath53 since one topological channel ( correlated to the lowest lls ) is canceled out at the n - n junction interfaces ( see fig .
[ fg : cleancond](d ) ) .
further , for @xmath62 , @xmath64 shows the same results as those reported from two - terminal conductance measurements@xcite .
however , @xmath56 is found to be non - zero only for @xmath60 because the transverse conductance across the junction can only be attributed to the existence of topological interface states and their interplay .
therefore , the existence of nonzero @xmath56 is direct evidence that this interplay occurs at the chern insulator junction .
moreover , while there is no plateau for @xmath60 , the conductance values vary depending on @xmath63 . it has been shown that roughness on both edges and interfaces leads to ballistic - ohmic transitions in the quantum hall conductance plateaus , by demonstrating that experimental observations in the ohmic limit are distinct from quantum hall conductance in the ballistic limit@xcite . here
, our approach to four - terminal conductance is in the ballistic limit , which necessitates the introduction of the roughness effects in order to treat the system more realistically .
edge roughness is randomly generated by @xmath65 where @xmath66.\end{aligned}\ ] ] here , @xmath67 and @xmath68 are independent random numbers within range @xmath69 $ ] , and the root mean square roughness is about 0.53 @xmath70 . and ( b ) @xmath71 as functions of @xmath63 .
the dashed lines are guiding lines corresponding to the individual plateaus , and the shaded regions represent different chern number configurations @xmath72 .
results are obtained by averaging over 100 randomly distributed edge roughness sets .
, width=321 ] figure [ fg : roughcond ] shows the effects of edge roughness on the four - terminal conductance of bipolar graphene in the quantum hall regime .
for @xmath60 , both @xmath55 and @xmath56 now exhibit clear plateaus .
the plateaus appear at @xmath53 , indicating an equal split into the two individual leads .
the roughness - induced conductance plateau can be intuitively understood by the valley - isospin dependent quantum hall conductance of armchair graphene nanoribbons@xcite .
we find that the resulting conductance plateau @xmath53 in fig .
[ fg : roughcond](a ) is equal to the mean of the conductance values for three different armchair graphene nanoribbon cases : @xmath73 for the metallic case , and @xmath74 for the two semiconducting cases@xcite . in other words ,
quantum hall conductance through a bipolar junction is expected to be three - fold , depending on the size of the armchair graphene nanoribbons .
as previously mentioned , randomly distributed edge roughness allows the system to contain those three different ribbon widths . additionally , the four - terminal conductance results satisfy flux conservation : the incoming flux from lead 1 is given by @xmath52 because there is a single metallic channel due to @xmath75 , and the sum of @xmath55 and @xmath56 is equal to @xmath52 , resulting in @xmath76 .
the half - and - half conductance splitting leads to the conclusion that the interplay of the topological interface states from the lowest lls undergoes full mixing .
compared to the lowest - ll case , the interplay nature between higher - index lls at the bipolar junction in quantum hall graphene is more complicated .
figure [ fg : roughcond](b ) shows the four - terminal conductance results in the presence of edge roughness for @xmath77 , where two topological edge modes exist in each region . note that the contribution of these topological states to the conductance is given by @xmath78 because the higher - index ll is doubly degenerate compared to the lowest ll . around @xmath1 it is seen that @xmath55 is twice as large as @xmath56 , satisfying flux conservation , i.e. , @xmath79 .
this asymmetric splitting behaviour can be explained by considering the chalker - coddington type argument , where it has been confirmed that the interplay of topological states at the interface can be expressed as a transfer matrix representation@xcite . in the case of @xmath77 , there are four modes in total with the corresponding transfer matrix reading@xcite @xmath80 where @xmath81 , @xmath82 .
here , @xmath83 indicates the degree of interplay between the lowest- ( 1st- ) ll topological interface states , and @xmath84 gives rise to the half - and - half mixing between the states . for @xmath85 , there is no interplay , whereas for @xmath86 , the incoming mode from one region completely transfers to the other region .
note that the matrix @xmath87 , which is composed of two diagonal block matrices , implies no coupling between the lowest- and 1st - ll topological states due to the orthogonality of the solutions ( see eq .
( [ eq : analsols ] ) ) . in other words
, the orthogonality allows for interplay between the time - reversal partners , based on the fact that particle - hole symmetry exists in the system .
as previously stated , the lowest - ll topological states undergo half - and - half mixing , resulting in @xmath88 , while on the other hand , @xmath89 is expected to be somewhere between 0 and @xmath90 . with the given @xmath88 , @xmath89
can be found using the transfer matrix with the following equation : @xmath91 where @xmath92 and @xmath93 .
note that , for @xmath94 , @xmath95 and @xmath96 are electron - type topological states corresponding to @xmath55 , and @xmath97 and @xmath98 are hole - type states corresponding to @xmath56 . for @xmath99 , @xmath89
is found to be about @xmath100 , resulting in @xmath101 and @xmath102 as shown in fig .
[ fg : roughcond](b ) .
consequently , the higher - ll topological states are expected to be `` moderately equilibrated '' at the chern insulator junction .
moreover , it is worth investigating the interplay behaviour for asymmetric chern number configurations , e.g. , @xmath103 and @xmath104 , as displayed in fig .
[ fg : roughcond](b ) . unlike the symmetric cases ,
higher - ll topological states have no time - reversal partner for interplay , differing from the lowest - ll states .
however , because of the asymmetry , the interplay between the lowest - ll states can be changed , making @xmath105 no longer true .
indeed , it is straightforward to find values of @xmath106 for various asymmetric configurations of chern numbers , based on the chalker - coddington type argument : @xmath107 for @xmath103 , @xmath108 for @xmath104 , and so on . as discussed above ,
the mixing nature of the topological interface states at the chern insulator junction is fairly complicated because a number of interface states can be coupled to each other like as a network , and behave according to the chalker - coddington argument . on the other hand ,
there have been findings where mode - mixing at the chern insulator junction can be limited to the lowest - ll mode in the presence of a `` depletion spacer''@xcite .
in fact , realistic device fabrication inevitably causes a smoothening of the potential step for the bipolar junction , which can be described by eq .
( [ eq : bipolarcontinuum ] ) . by varying the parameter @xmath109 from the potential profiles ( eq .
( [ eq : bipolarcontinuum ] ) ) , the gradualness of the smoothly varying potential profile can be tuned .
as expected , sufficiently large values of @xmath109 lead to a reduction in @xmath56 , ending up with clear plateaus at @xmath53 regardless of the chern number configurations , as shown in fig .
[ fg : suppression_mixing](a ) .
the reduced conductance from lead 1 to 4 implies that the interplay between topological states is limited to the lowest lls as recently discussed@xcite . in other words ,
suppression of the interplay occurs only for the higher - index lls .
an intuitive explanation for this suppression phenomena can be found by analyzing the probability density when the potential profiles become gradual .
figures [ fg : suppression_mixing](b ) and ( c ) provide a comparison of the probability density of the scattering region between the abrupt and gradual potential profiles .
it is clearly seen that the topological interface states are spatially impelled toward the interior of the system , so that the degree of the interplay at the interface decreases .
it is speculated that the lowest - ll topological states are formed at @xmath50 , where the potential strength is exactly zero , so that the topological states at @xmath50 are almost unaffected by the gradual potential profile .
higher - index ll topological states though are further separated by the gradually varying potentials .
regarding the mode - mixing suppression , it is natural to expect that the suppression of the interplay between the topological interface states depends on the gradualness of the junction profile .
indeed , figs .
[ fg : suppression_mixing](d ) and ( e ) confirm that @xmath56 for the @xmath110 case reaches @xmath53 more rapidly than for the @xmath111 case , as @xmath109 varies .
as expected , the interplay ends up being attributed to the lowest - ll states only , for the sufficiently gradual potential profiles .
therefore , a careful analysis of the conductance plateaus is required to correctly understand the quantum hall measurements through bipolar junctions , including an examination of potential profile steepness .
depending on the slope of the potential profile near the interface , the degree of interplay between topological states can vary , likely resulting in unclear plateau values .
in summary , we have exploited four - terminal quantum hall conductance through a bipolar junction in graphene to understand the interplay nature of the topologically protected states at the junction interface .
the properties of the topological interface states are examined not only by numerical calculations but also analytic solutions , which are in agreement with each other .
we have studied the influence of edge roughness on the splitting of the quantum hall conductance due to the interplay between the topological interface states . from the resulting quantum hall conductance plateaus ,
we have deduced the degree of interplay for various chern number configurations using the chalker - coddington type argument .
we conclude by remarking that , in general , the interplay between topological interface states is described by partial equilibration at a clean interface , as distinct from full equilibration at a rough interface .
moreover , for realistic bipolar junctions in graphene , interplay between higher - ll topological states tends to be suppressed as the potential profiles become gradual .
such a suppression of the interplay phenomena occurs because the higher - ll topological states become spatially farther away from the interface for gradually varying potentials .
our findings in this study provide a good predictive way of elucidating the interplay behaviour of quantum hall graphene systems .
this work was sponsored by project code ( ibs - r024-d1 ) .
nm set up the problem and the corresponding system for both analytical and numerical studies .
nm carried out the calculations and analyse the calculated results upon discussion with hcp .
nm and hcp wrote the manuscript .
40 k. s. novoselov , a. k. geim , s. v. morozov , d. jiang , m. i. katsnelson , i. v. grigorieva , s. v. dubonos and a. a. two - dimensional gas of massless dirac fermions in graphene _ nature ( london ) _ * 438 * , 197 - 200 ( 2005 ) .
a. h. castro neto , f. guinea , n. m. r. peres , k. s. novoselov and a. k. geim the electronic properties of graphene _ rev .
* 81 * , 109 ( 2009 ) .
j. r. williams , l. dicarlo and c. m. marcus quantum hall effect in a gate - controlled _ p - n _ junction of graphene _ science _ * 317 * , 638 - 641 ( 2007 ) .
d. a. abanin and l. s. levitov quantized transport in graphene _
p - n _ junctions in a magnetic field _ science _ * 317 * , 641 - 643 ( 2007 ) .
y. hatsugai chern number and edge states in the integer quantum hall effect _ phys .
lett . _ * 71 * , 3697 ( 1993 ) . c. l. kane and e. j. mele z@xmath112 topological order and the quantum spin hall effect _ phys .
lett . _ * 95 * , 146802 ( 2005 ) .
b. i. halperin quantized hall conductance , current - carrying edge states , and the existence of extended states in a two - dimensional disordered potential _ phys .
b _ * 25 * , 2185 ( 1982 ) . a. m. essin and v. gurarie bulk - boundary correspondence of topological insulators from their respective green s functions _ phys . rev . b _ * 84 * , 125132 ( 2011 ) .
s. matsuura , p .- y .
chang , a. p. schnyder and s. ryu protected boundary states in gapless topological phases _ new j. phys .
_ * 15 * 065001 ( 2013 ) .
d. j. thouless , m. kohmoto , m. p. nightingale and m. quantized hall conductance in a two - dimensional periodic potential _ phys
lett . _ * 49 * 405 ( 1982 ) .
f. d. m. haldane model for a quantum hall effect without landau levels : condensed - matter realization of the `` parity anomaly '' _ phys .
lett . _ * 61 * , 2015 ( 1988 ) .
m. z. hasan and c. l. kane colloquium : topological insulators _ rev .
phys . _ * 82 * , 3045 ( 2011 ) .
y. ando topological insulator materials _ j. phys .
* 82 * , 102001 ( 2013 ) . j. li and s .- q .
shen disorder effects in the quantum hall effect of graphene _
p - n _ junctions , _ phys .
b _ * 78 * , 205308 ( 2008 ) .
t. low ballistic - ohmic quantum hall plateau transition in a graphene _ p - n _ junction _ phys .
b _ * 80 * , 205423 ( 2009 ) . c. frdorf , l. trifunovic , n. bogdanoff and p. w. brouwer graphene _ pn _ junction in a quantizing magnetic field : conductance at intermediate disorder strength _ phys . rev . b _ * 94 * , 195439 ( 2016 ) . s. w. lagasse and j. u. lee theory of landau level mixing in heavily graded graphene _
p - n _ junctions _ phys .
b _ * 94 * , 165312 ( 2016 ) .
ki and h .- j .
lee quantum hall resistances of a multiterminal top - gated graphene device _ phys .
b _ * 79 * , 195327 ( 2009 ) .
s. debey and m. m. deshmukh tuning equilibration of quantum hall edge states in graphene - role of crossed electric and magnetic fields _ solid state commun . _
* 237 - 238 * , 59 - 63 ( 2016 ) .
nam , h .- j .
lee and b. zylmaz dependence of quantum - hall conductance on the edge - state equilibration position in a bipolar graphene sheet _ phys .
b _ * 81 * , 033301 ( 2010 ) .
b. zylmaz , p. jarillo - herrero , d. efetov , d. a. abanin , l. s. levitov and p. kim electronic transport and quantum hall effect in bipolar graphene _
p - n - p _ junctions _ phys . rev .
lett . _ * 99 * , 166804 ( 2007 ) .
l. oroszny , p. rakyta , a. kormnyos , c. j. lambert and j. cserti theory of snake states in graphene _ phys .
b _ * 77 * , 081403(r ) ( 2008 ) .
s. park and h .- s .
sim magnetic edge states in graphene in nonuniform magnetic fields _ phys . rev .
b _ * 77 * , 075433 ( 2008 )
. t. k. ghosh , a. de martino , w. hssler , l. dellanna and r. egger conductance quantization and snake states in graphene magnetic waveguides _ phys .
b _ * 77 * , 081404(r ) ( 2008 ) . l. cohnitz , a. de martino , w. hssler and r. egger chiral interface states in graphene _
p - n _ junctions _ phys .
b _ * 94 * , 165433 ( 2016 ) . c. w. groth , m. wimmer , a. r. akhmerov and x. waintal kwant : a software package for quantum transport _ new j. phys . _
* 16 * , 063065 ( 2014 ) .
j. tworzydo , i. snynam , a. r. akhmerov and c. w. beenakker valley - isospin dependence of the quantum hall effect in a graphene _ p - n _ junction _ phys .
b _ * 76 * , 035411 ( 2007 ) .
a. mreca - kolasiska , s. heun and b. szafran aharonov - bohm interferometer based on _
n - p _ junctions in graphene nanoribbons _ phys . rev .
b _ * 93 * , 125411 ( 2016 ) .
k. nakada , m. fujita , g. dresselhaus and m. s. dresselhaus edge state in graphene ribbons : nanometer size effect and edge shape dependence _ phys
b _ _ 54 _ , 17954 ( 1996 ) .
y .- w . son , m. l. cohen and s. g. louie energy gaps in graphene nanoribbons _ phys . rev .
lett . _ * 97 * , 216803 ( 2006 ) . j. t. chalker and p. d. coddington percolation ,
quantum tunneling and the integer hall effect _ j. phys . : condens .
_ * 21 * , 2665 - 2679 ( 1988 )
lee , z. wang and s. kivelson quantum percolation and plateau transitions in the quantum hall effect _ phys .
lett . _ * 70 * , 4130 ( 1993 ) . | we investigate the interplay behaviour of the topologically protected states at the interface of chern insulator junctions generated by quantum hall graphene systems .
four - terminal quantum hall conductance is exploited to understand how the states are coupled to each other at the interface .
we demonstrate that the interplay of the topological states occurs only between the time - reversal partners characterized by the same chern number , using the chalker - coddington type argument .
we also find interplay suppression , depending on how the chern insulator junction is fabricated . |
cellular systems are evolving into heterogeneous networks consisting of distributed base stations ( bss ) covering overlapping areas of different sizes , and thus the problems of interference management and cell association are becoming complicated and challenging @xcite .
one of the most promising solutions to these problems is given by so called _ cloud radio access networks _
, in which the encoding / decoding functionalities of the bss are migrated to a central unit .
this is done by operating the bss as `` soft '' relays that interface with the central unit via backhaul links used to carry only baseband signals ( and not `` hard '' data information ) @xcite-@xcite .
cloud radio access networks are expected not only to effectively handle the inter - cell interference but also to lower system cost related to the deployment and management of the bss .
however , one of the main impairments to the implementation of cloud radio access networks is given by the capacity limitations of the digital backhaul links connecting the bss and the central unit .
these limitations are especially pronounced for pico / femto - bss , whose connectivity is often afforded by last - mile cables @xcite@xcite , and for bss using wireless backhaul links @xcite . in the _ uplink _ of cloud radio access networks , each bs compresses its received signal to the central unit via its finite - capacity backhaul link .
the central unit then performs joint decoding of all the mobile stations ( mss ) based on all received compressed signals .
recent theoretical results have shown that _ distributed compression _
schemes @xcite can provide significant advantages over the conventional approach based on independent compression at the bss .
this is because the signals received by different bss are statistically correlated @xcite-@xcite , and hence distributed source coding enables the quality of the compressed signal received from one bs to be improved by leveraging the signals received from the other bss as side information .
note that the correlation among the signals received by the bss is particularly pronounced for systems with many small cells concentrated in given areas .
while current implementations @xcite@xcite employ conventional independent compression across the bss , the advantages of distributed source coding were first demonstrated in @xcite , and then studied in more general settings in @xcite-@xcite .
related works based on the idea of computing a function of the transmitted codewords at the bss , also known as compute - and - forward , can be found in @xcite@xcite . in the _ downlink _ of cloud radio access networks ,
the central encoder performs joint encoding of the messages intended for the mss .
then , it independently compresses the produced baseband signals to be transmitted by each bs .
these baseband signals are delivered via the backhaul links to the corresponding bss , which simply upconvert and transmit them through their antennas .
this system was studied in @xcite@xcite . in particular , in @xcite ,
the central encoder performs dirty - paper coding ( dpc ) @xcite of all mss signals before compression .
a similar approach was studied in @xcite by accounting for the effect of imperfect channel state information ( csi ) .
reference @xcite instead proposes strategies based on compute - and - forward , showing advantages in the low - backhaul capacity regime and high sensitivity of the performance to the channel parameters . for a review of more conventional strategies in which the backhaul links are used to convey message information , rather than the compressed baseband signals ,
we refer to @xcite-@xcite .
in this work , we propose a novel approach for the compression on the backhaul links of cloud radio access networks in the downlink that can be seen as the counterpart of the distributed source coding strategy studied in @xcite-@xcite for the uplink .
moreover , we propose the joint design of precoding and compression .
a key idea is that of allowing the quantization noise signals corresponding to different bss to be correlated with each other . the motivation behind this choice is the fact that a proper design of the correlation of the quantization noises across the bss can be beneficial in limiting the effect of the resulting quantization noise seen at the mss .
in order to create such correlation , we propose to jointly compress the baseband signals to be delivered over the backhaul links using so called _ multivariate compression _
* ch . 9 ) .
we also show that , in practice , multivariate compression can be implemented without resorting to joint compression across all bss , but using instead a successive compression strategy based on a sequence of minimum mean squared error ( mmse ) estimation and per - bs compression steps . after reviewing some preliminaries on multivariate compression in sec .
[ sec : preliminaries ] , we formulate the problem of jointly optimizing the precoding matrix and the correlation matrix of the quantization noises with the aim of maximizing the weighted sum - rate subject to power and the backhaul constraints resulting from multivariate compression in sec .
[ sec : problem formulation ] .
there , we also introduce the proposed architecture based on successive per - bs steps .
we then provide an iterative algorithm that achieves a stationary point of the problem in sec .
[ sec : joint ] .
moreover , we compare the proposed joint design with the more conventional approaches based on independent backhaul compression @xcite-@xcite or on the separate design of precoding and ( multivariate ) quantization in sec .
[ sec : separate ] .
the robust design with respect to imperfect csi is also discussed in detail . in sec .
[ sec : numerical - results ] , extensive numerical results are provided to illustrate the advantages offered by the proposed approach .
the paper is terminated with the conclusion in sec .
[ sec : conclusions ]
. _ notation _ : we adopt standard information - theoretic definitions for the mutual information @xmath0 between the random variables @xmath1 and @xmath2 , conditional mutual information @xmath3 between @xmath1 and @xmath2 conditioned on random variable @xmath4 , differential entropy @xmath5 of @xmath1 and conditional differential entropy @xmath6 of @xmath1 conditioned on @xmath2 @xcite .
the distribution of a random variable @xmath1 is denoted by @xmath7 and the conditional distribution of @xmath1 conditioned on @xmath2 is represented by @xmath8 .
all logarithms are in base two unless specified . the circularly symmetric complex gaussian distribution with mean @xmath9 and covariance matrix @xmath10 is denoted by @xmath11 .
the set of all @xmath12 complex matrices is denoted by @xmath13 , and @xmath14 represents the expectation operator .
we use the notations @xmath15 and @xmath16 to indicate that the matrix @xmath17 is positive semidefinite and positive definite , respectively . given a sequence @xmath18
, we define a set @xmath19 for a subset @xmath20 .
the operation @xmath21 denotes hermitian transpose of a matrix or vector , and notation @xmath22 is used for the correlation matrix of random vector @xmath23 , i.e. , @xmath24 $ ] ; @xmath25 represents the cross - correlation matrix @xmath26 $ ] ; @xmath27 is used for the conditional correlation matrix , i.e. , @xmath28 $ ] .
downlink communication in a cloud radio access network in which there are @xmath29 multi - antenna bss and @xmath30 multi - antenna mss .
the @xmath29 bss include both macro - bss and pico / femto - bss .
the @xmath30 mss are distributed across all the cells.,width=566,height=434 ]
we consider the downlink of a cloud radio access network as illustrated in fig .
[ fig : block diagram ] . in the system
, a central encoder communicates to @xmath30 mss through @xmath29 distributed bss .
the message @xmath31 for each @xmath32th ms is uniformly distributed in the set @xmath33 , where @xmath34 is blocklength and @xmath35 is the information rate of message @xmath31 in bits per channel use ( c.u . ) .
each ms @xmath32 has @xmath36 receive antennas for @xmath37 , and each bs @xmath38 is equipped with @xmath39 antennas for @xmath40 .
note that the bss can be either macro - bss or pico / femto - bss and that the mss are arbitrarily distributed across the cells .
each @xmath38th bs is connected to the central encoder via digital backhaul link with finite - capacity @xmath41 bits per c.u . for notational convenience ,
we define @xmath42 as the total number of transmitting antennas , @xmath43 as the total number of receive antennas , and the sets @xmath44 and @xmath45 . as shown in fig .
[ fig : central encoder ] , each message @xmath31 is first encoded by a separate channel encoder , which produces a coded signal @xmath46 .
the signal @xmath47 corresponds to the @xmath48 vector of encoded symbols intended for the @xmath32th ms for a given c.u . , and we have @xmath49 .
we assume that each coded symbol @xmath46 is taken from a conventional gaussian codebook so that we have @xmath50 .
the signals @xmath51 are further processed by the central encoder in two stages , namely _ precoding _ and _ compression_. as is standard practice , precoding is used in order to control the interference between the data streams intended for the same ms and for different mss . instead , compression is needed in order to produce the @xmath29 rate - limited bit streams delivered to each bs over the corresponding backhaul link .
specifically , recall that each bs @xmath38 receives up to @xmath41 bits per c.u . on the backhaul link from the central encoder .
further discussion on precoding and compression can be found in sec .
[ sec : problem formulation ] . on the basis of the bits received on the backhaul links
, each bs @xmath38 produces a vector @xmath52 for each c.u . , which is the baseband signal to be transmitted from its @xmath39 antennas .
we have the per - bs power constraints\leq\delta_{l}$ ] for @xmath53 , where the matrix @xmath54 is a non - negative definite matrix ( see , e.g. , ( * ? ? ?
ii - c ) ) . ]
@xmath55\leq p_{i},\,\,\mathrm{for}\,\ , i\in\mathcal{n_{b}}.\label{eq : perbs power constraint}\ ] ] assuming flat - fading channels , the signal @xmath56 received by ms @xmath32 is written as @xmath57 where we have defined the aggregate transmit signal vector @xmath58^{\dagger}$ ] , the additive noise @xmath59 to obtain ( [ eq : received signal ms ] ) .
] , and the channel matrix @xmath60 toward ms @xmath32 as @xmath61,\label{eq : channel to ms k}\ ] ] with @xmath62 denoting the channel matrix from bs @xmath38 to ms @xmath32 .
the channel matrices remain constant for the entire coding block duration .
we assume that the central encoder has information about the global channel matrices @xmath63 for all @xmath64 and that each ms @xmath32 is only aware of the channel matrix @xmath63 .
the bss must also be informed about the compression codebooks used by the central encoder , as further detailed later .
the case of imperfect csi at the central encoder will be discussed in sec .
[ sub : robust - design - with ] . based on the definition given above and assuming single - user detection at each ms , the rates @xmath65 can be achieved for each ms @xmath64 .
this section reviews some basic information - theoretical results concerning multivariate compression , which will be leveraged in the analysis of the proposed backhaul compression strategy in sec . [
sub : multivariate - backhaul - compression ] .
illustration of ( _ a _ ) conventional compression ; ( _ b _ ) multivariate compression.,width=457,height=289 ] to fix the ideas and the notation , we first consider the conventional compression problem illustrated in fig .
[ fig : compression examples]-(a ) .
the compression unit compresses a random sequence @xmath66 of @xmath34 independent and identically distributed ( i.i.d . ) samples with distribution @xmath7 at a rate of @xmath67 bits per symbol .
specifically , the compressor selects a codeword @xmath68 within a codebook @xmath69 of size @xmath70 and sends the corresponding index , of @xmath71 bits , to the decompression unit . at the decompression unit
, the sequence @xmath72 indicated by the received index is recovered . using the standard information - theoretic formulation , the compression strategy
is specified by a conditional distribution @xmath73 , which is referred to as the _ test channel _ ( see ,
e.g. , ( * ? ? ?
2 ) ) . for a given test channel ,
compression consists in selecting a sequence @xmath72 that is jointly typical and @xmath74 are called jointly typical with respect to a distribution @xmath75 if their joint empirical distribution ( i.e. , normalized histogram with step size @xmath76 ) does not deviate much from @xmath75 ( see , e.g. , ( * ? ? ?
2 ) for a formal definition ) . ] with the sequence @xmath66 with respect to the given joint distribution @xmath77 .
compression is hence successful if the encoder is able to find a jointly typical sequence @xmath68 in the codebook @xmath69 .
a classical result in information theory is that this happens with arbitrarily large probability as the block length @xmath34 grows large if the inequality @xmath78 is satisfied ( * ? ? ?
* ch . 3)@xcite .
we now review the more general multivariate compression illustrated in fig .
[ fig : compression examples]-(b ) . here , the sequence @xmath66 is compressed into @xmath79 indices with the goal of producing correlated compressed versions @xmath80 .
each @xmath38th index indicates a codeword within a codebook @xmath81 of size @xmath82 , and is sent to the @xmath38th decompression unit for @xmath83 .
each @xmath38th decompression unit then recovers a sequence @xmath84 corresponding to the received index .
we emphasize that the choice of the codewords @xmath80 is done jointly at the compression unit .
in particular , the specification of the compression strategy is given by a test channel @xmath85 .
this implies that the compression unit wishes to find codewords @xmath80 that are jointly typical with the sequence @xmath66 with respect to the given joint distribution @xmath86 .
the following lemma provides a sufficient condition for multivariate compression to be successful ( we refer to ( * ? ? ?
* lemma 8.2 ) for a more precise statement ) .
[ lem : multivariate covering ] _ ( * ? ? ?
9 ) _ consider an i.i.d .
sequence @xmath66 and @xmath34 large enough .
then , there exist codebooks @xmath87 with rates @xmath88 , that have at least one tuple of codewords @xmath89 jointly typical with @xmath66 with respect to the given joint distribution @xmath86 with probability arbitrarily close to one , if the inequalities @xmath90 are satisfied .
see ( * ? ? ?
9 ) for a proof . we observe that , for a given test channel @xmath85 , the inequalities ( [ eq : multivariate covering condition ] ) impose joint conditions on the rate of all codebooks .
this is due to the requirement of finding codewords @xmath91 that are jointly correlated according to the given test channel @xmath85 .
also , note that the vector @xmath66 may be such that each @xmath92 is itself a vector and that the distortion requirements at each decompression unit prescribe that the decompression unit be interested in only a subset of entries in this vector .
the connection between the multivariate set - up in fig .
[ fig : compression examples]-(b ) and the system model under study in fig .
[ fig : block diagram ] will be detailed in the next section .
in this section , we first propose a novel precoding - compression strategy based on multivariate compression for the downlink of a cloud radio access network .
we then establish the problem definition . finally , a novel architecture that implements multivariate compression via a sequence of mmse estimation and per - bs compression steps is proposed .
illustration of the operation at the central encoder.,width=642,height=283 ] as mentioned in the previous section , the operation at the central encoder can be represented by the block diagram in fig .
[ fig : central encoder ] . specifically , after channel encoding , the encoded signals @xmath93^{\dagger}$ ] undergo precoding and compression , as detailed next .
* 1 . precoding : * in order to allow for interference management both across the mss and among the data streams for the same ms ,
the signals in vector @xmath94 are linearly precoded via multiplication of a complex matrix @xmath95 .
the precoded data can be written as @xmath96 where the matrix @xmath97 can be factorized as @xmath98,\label{eq : whole beamformer}\ ] ] with @xmath99 denoting the precoding matrix corresponding to ms @xmath32 .
the precoded data @xmath100 can be written as @xmath101^{\dagger}$ ] , where the signal @xmath102 is the @xmath103 precoded vector corresponding to the @xmath38th bs and given as @xmath104 with the matrix @xmath105 having all zero elements except for the rows from @xmath106 to @xmath107 which contain an @xmath108 identity matrix .
note that non - linear precoding using dpc techniques can also be considered , as discussed in remark [ rem : dpc ] below .
compression : * each precoded data stream @xmath102 for @xmath109 must be compressed in order to allow the central encoder to deliver it to the @xmath38th bs through the backhaul link of capacity @xmath41 bits per c.u .
each @xmath38th bs then simply forwards the compressed signal @xmath110 obtained from the central encoder .
note that this implies that the bss need not be aware of the channel codebooks and of the precoding matrix @xmath97 used by the central encoder .
instead , they must be informed about the quantization codebooks selected by the central encoder . using standard rate - distortion considerations
, we adopt a gaussian test channel to model the effect of compression on the backhaul link .
in particular , we write the compressed signals @xmath110 to be transmitted by bs @xmath38 as is seemingly more general than ( [ eq : gaussian test channel each bs ] ) , but this can be captured by adjusting the matrix @xmath97 in ( [ eq : precoding only ] ) . ]
@xmath111 where the compression noise @xmath112 is modeled as a complex gaussian vector distributed as @xmath113 .
overall , the vector @xmath58^{\dagger}$ ] of compressed signals for all the bss is given by @xmath114 where the compression noise @xmath115^{\dagger}$ ] is modeled as a complex gaussian vector distributed as @xmath116 .
the compression covariance @xmath117 is given as @xmath118,\label{eq : compression covariance}\ ] ] where the matrix @xmath119 is defined as @xmath120 $ ] and defines the correlation between the quantization noises of bs @xmath38 and bs @xmath121 .
rate - distortion theory guarantees that compression codebooks can be found for any given covariance matrix @xmath122 under appropriate constraints imposed on the backhaul links capacities .
this aspect will be further discussed in sec .
[ sub : multivariate - compression theory ] . with the described precoding and compression operations ,
the achievable rate ( [ eq : rate ms k ] ) for ms @xmath32 is computed as @xmath123 [ rem : compression]in the conventional approach studied in @xcite-@xcite , the signals @xmath102 corresponding to each bs @xmath38 are compressed independently .
this corresponds to setting @xmath124 for all @xmath125 in ( [ eq : compression covariance ] ) .
a key contribution of this work is the proposal to leverage correlated compression for the signals of different bss in order to better control the effect of the additive quantization noises at the mss .
the design of the precoding matrix @xmath97 and of the quantization covariance @xmath117 can be either performed separately , e.g. , by using a conventional precoder @xmath97 such as zero - forcing ( zf ) or mmse precoding ( see , e.g. , @xcite-@xcite ) , or jointly .
both approaches will be investigated in the following .
[ rem : dpc]if non - linear precoding via dpc @xcite is deployed at the central encoder with a specific encoding permutation @xmath126 of the ms indices @xmath127 , the achievable rate @xmath128 for ms @xmath129 is given as @xmath130 in lieu of ( [ eq : rate ms k ] ) and can be calculated as @xmath131 with the function @xmath132 given as @xmath133 note that the dpc is designed based on the knowledge of the noise levels ( including the quantization noise ) in order to properly select the mmse scaling factor @xcite . as explained above , due to the fact that the bss are connected to the central encoder via finite - capacity backhaul links ,
the precoded signals @xmath102 in ( [ eq : precoding bs - wise ] ) for @xmath109 are compressed before they are communicated to the bss using the gaussian test channels ( [ eq : gaussian test channel each bs ] ) . in the conventional case
in which the compression noise signals related to the different bss are uncorrelated , i.e. , @xmath124 for all @xmath134 as in @xcite-@xcite , the signal @xmath110 to be emitted from bs @xmath38 can be reliably communicated from the central encoder to bs @xmath38 if the condition @xmath135 is satisfied for @xmath109 .
this follows from standard rate - distortion theoretic arguments ( see , e.g. , @xcite and sec .
[ sub : compression ] ) .
we emphasize that ( [ eq : independent comp condition ] ) is valid under the assumption that each bs @xmath38 is informed about the quantization codebook used by the central encoder , as defined by the covariance matrix @xmath136 . in this paper , we instead propose to introduce correlation among the compression noise signals , i.e. , to set @xmath137 for @xmath125 , in order to control the effect of the quantization noise at the mss . as discussed in sec .
[ sec : preliminaries ] , introducing correlated quantization noises calls for joint , and not independent , compression of the precoded signals of different bss .
as seen , the family of compression strategies that produce descriptions with correlated compression noises is often referred to as _ multivariate compression .
_ by choosing the test channel according to ( [ eq : whole encoding operation ] ) ( see sec . [ sub : multivariate - compression theory ] ) , we can leverage lemma [ lem : multivariate covering ] to obtain sufficient conditions for the signal @xmath110 to be reliably delivered to bs @xmath38 for all @xmath109 . in lemma
[ lem : multivariate ] , we use @xmath138 to denote the matrix obtained by stacking the matrices @xmath139 for @xmath140 horizontally .
[ lem : multivariate ] the signals @xmath141 obtained via the test channel ( [ eq : whole encoding operation ] ) can be reliably transferred to the bss on the backhaul links if the condition @xmath142 is satisfied for all subsets @xmath143 .
the proof follows by applying lemma [ lem : multivariate covering ] by substituting @xmath144 for the signal @xmath1 to be compressed , and @xmath141 for the compressed versions @xmath91 . comparing ( [ eq : independent comp condition ] ) with
( [ eq : multivariate computed ] ) shows that the introduction of correlation among the quantization noises for different bss leads to additional constraints on the backhaul link capacities .
assuming the operation at the central encoder , bss and mss detailed above , we are interested in maximizing the weighted sum - rate @xmath145 subject to the backhaul constraints ( [ eq : multivariate computed ] ) over the precoding matrix @xmath97 and the compression noise covariance @xmath117 for given weights @xmath146 , @xmath64
. this problem is formulated as [ eq : original problem ] @xmath147 the condition ( [ eq : original problem backhaul ] ) corresponds to the backhaul constraints due to multivariate compression as introduced in lemma [ lem : multivariate ] , and the condition ( [ eq : original problem power ] ) imposes the transmit power constraints ( [ eq : perbs power constraint ] ) .
it is noted that the problem ( [ eq : original problem ] ) is not easy to solve due to the non - convexity of the objective function @xmath148 in ( [ eq : original problem objective ] ) and the functions @xmath149 in ( [ eq : original problem backhaul ] ) with respect to @xmath150 . in sec . [
sec : joint ] , we will propose an algorithm to tackle the solution of problem ( [ eq : original problem ] ) . proposed architecture for multivariate compression based on successive steps of mmse estimation and per - bs compression.,width=623,height=347 ] in order to obtain correlated quantization noises across bss using multivariate compression
, it is in principle necessary to perform joint compression of all the precoded signals @xmath102 corresponding to all bss @xmath38 for @xmath109 ( see sec . [ sub : multivariate - compression theory ] ) .
if the number of bss is large , this may easily prove to be impractical .
here , we argue that , in practice , joint compression is not necessary and that the successive strategy based on mmse estimation and per - bs compression illustrated in fig .
[ fig : successive comp ] is sufficient .
the proposed approach works with a fixed permutation @xmath151 of the bss indices @xmath152 .
the central encoder first compresses the signal @xmath153 using the test channel ( [ eq : gaussian test channel each bs ] ) , namely @xmath154 , with @xmath155 , and sends the bit stream describing the compressed signal @xmath156 over the backhaul link to bs @xmath157 .
then , for any other @xmath109 with @xmath158 , the central encoder obtains the compressed signal @xmath159 for bs @xmath160 in a successive manner in the given order @xmath161 by performing the following steps : * ( _ a _ ) * * _ estimation _ : * the central encoder obtains the mmse estimate @xmath162 of @xmath159 given the signal @xmath163 and the previously obtained compressed signals @xmath164 .
this estimate is given by @xmath165\label{eq : mmse estimate x_hat_i}\\ & = \mathbf{\sigma}_{\mathbf{x}_{\pi(i)},\mathbf{u}_{\pi(i)}}\mathbf{\sigma}_{\mathbf{u}_{\pi(i)}}^{-1}\mathbf{u}_{\pi(i)},\nonumber\end{aligned}\ ] ] where we defined the vector @xmath166^{\dagger}$ ] , and the correlation matrices @xmath167 and @xmath168 are given as @xmath169\end{aligned}\ ] ] and @xmath170,\end{aligned}\ ] ] with @xmath171 for subsets @xmath172 and the set @xmath173 defined as @xmath174 .
* ( _ b _ ) * * _ compression _ : * the central encoder compresses the mmse estimate @xmath175 to obtain @xmath159 using the test channel @xmath176 where the quantization noise @xmath177 is independent of the estimate @xmath175 and distributed as @xmath178 with @xmath179 note that the first equality in ( [ eq : conditional covariance ] ) follows from the fact that the mmse estimate @xmath175 is a sufficient statistic for the estimation of @xmath159 from @xmath180 ( see , e.g. , @xcite ) .
moreover , the compression rate @xmath181 required by the test channel ( [ eq : compression after mmse ] ) is given by @xmath182 to see why the structure in fig .
[ fig : successive comp ] described above realizes multivariate compression , we need the following lemma . [
lem : contra polymatroid ] the region of the backhaul capacity tuples @xmath183 satisfying the constraints ( [ eq : original problem backhaul ] ) is a _ _ contrapolymatroid _ _ and @xmath184 with @xmath185 being the set of all subsets of @xmath186 .
then , the polyhedron @xmath187 is a contrapolymatroid if the function @xmath188 satisfies the conditions : ( _ a _ ) @xmath189 ; ( _ b _ ) @xmath190 if @xmath191 ; ( _ c _ ) @xmath192 ( * ? ? ?
therefore , it has a corner point for each permutation @xmath161 of the bs indices @xmath152 , and each such corner point is given by the tuple @xmath193 with @xmath194 for @xmath40 .
moreover , the corner point @xmath193 in ( [ eq : corner points ] ) is such that the constraints ( [ eq : original problem backhaul ] ) are satisfied with equality for the subsets @xmath195 @xmath196 .
this lemma follows immediately by the definition and properties of contrapolymatroids as summarized in ( * ? ? ?
moreover , the second equality of ( [ eq : corner points ] ) holds due to the fact that the mmse estimate @xmath175 is a sufficient statistic for the estimation of @xmath159 from @xmath180 ( see , e.g. , @xcite ) , or equivalently from the markov chain @xmath197 .
illustrative example of the contrapolymatroidal region of backhaul capacities @xmath198 satisfying the constraint ( [ eq : original problem backhaul ] ) imposed by multivariate compression for @xmath199 .
the two corner points are given by ( [ eq : corner points]).,width=453,height=347 ] lemma [ lem : contra polymatroid ] shows that the region of backhaul capacities that guarantees correct delivery of the compressed signals ( [ eq : gaussian test channel each bs ] ) to the bss , as identified in lemma [ lem : multivariate ] , is a type of polyhedron known as contrapolymatroid , as exemplified in fig .
[ fig : corner points ] .
a specific feature of contrapolymatroid is that the corner points can be easily characterized as in ( [ eq : corner points ] ) . from standard rate - distortion theory arguments
, the equality between ( [ eq : compression rate mmse compression ] ) and ( [ eq : corner points ] ) implies that the corner point @xmath193 can be obtained for any permutation @xmath161 by the successive estimation - compression strategy outlined above and illustrated in fig .
[ fig : successive comp ] .
overall , the discussion above shows that , for any correlation matrix @xmath117 in ( [ eq : compression covariance ] ) , multivariate compression is feasible by using the successive estimation - compression architecture in fig .
[ fig : successive comp ] if the backhaul capacities satisfy the corner point condition ( [ eq : corner points ] ) for the given bs order @xmath161 .
note that , in general , conditions ( [ eq : corner points ] ) are more restrictive than ( [ eq : original problem backhaul ] ) , which allows for any backhaul capacities in the contrapolymatroid .
this is because the solution to the optimization problem ( [ eq : original problem ] ) can be seen by contradiction to lie necessarily on the boundary of the region ( [ eq : original problem backhaul ] ) but possibly not on the corner points .
further discussion on this point can be found in sec .
[ sub : practical - implementation ] , where we observe that , in practice , this limitation is not critical .
in this section , we aim at jointly optimizing the precoding matrix @xmath97 and the compression covariance @xmath117 by solving problem ( [ eq : original problem ] ) . in sec .
[ sec : separate ] , we will then consider the generally suboptimal strategy in which the precoding matrix is fixed according to standard techniques , such as zf @xcite , mmse @xcite or weighted sum - rate maximizing precoding @xcite@xcite by neglecting the compression noise , and only the compression noise matrix @xmath117 is optimized . as mentioned , the optimization ( [ eq : original problem ] ) is a non - convex problem . to tackle this issue
, we will now show how to obtain an efficient scheme that is able to achieve a stationary point of problem ( [ eq : original problem ] ) . to this end
, we first make a change of variable by defining the variables @xmath200 for @xmath64 .
then , we define the functions @xmath201 and @xmath202 with respect to the variables @xmath203 which are obtained by substituting @xmath204 into the functions @xmath205 and @xmath149 in problem ( [ eq : original problem ] ) , respectively , and the transmit power constraint as @xmath206 for @xmath109 .
the so - obtained problem over the variables @xmath203 and @xmath117 is still non - convex due to the second term in @xmath201 and the first term in @xmath202 , which are concave in the variables @xmath203 and @xmath117 .
however , we observe that it falls into the class of difference - of - convex ( dc ) problems @xcite . among various algorithms having desirable properties for the solution of dc problems @xcite , we adopt the majorization minimization ( mm ) scheme @xcite , which solves a sequence of convex problems obtained by linearizing non - convex parts in the objective function @xmath201 and the constraint function @xmath202 .
it is shown that the mm algorithm converges to a stationary point of the original non - convex problems ( see , e.g. , ( * ? ? ?
* theorem 1 ) , ( * ? ? ?
1.3.3 ) and ( * ? ?
* theorem 3 ) ) .
the proposed algorithm is summarized in table algorithm 1 where we define the functions @xmath207 and @xmath208 as @xmath209 and @xmath210 with the function @xmath211 given as @xmath212 \1 .
initialize the matrices @xmath213 and @xmath214 to arbitrary feasible positive semidefinite matrices for problem ( [ eq : original problem ] ) and set @xmath215 .
update the matrices @xmath216 and @xmath217 as a solution of the following ( convex ) problem . @xmath218
\3 . go to step 4 if a convergence criterion is satisfied .
otherwise , set @xmath219 and go back to step 2 .
calculate the precoding matrices @xmath220 for @xmath64 , where @xmath221 is a diagonal matrix whose diagonal elements are the nonzero eigenvalues of @xmath222 and the columns of @xmath223 are the corresponding eigenvectors . as we have discussed in sec .
[ sec : problem formulation ] , given the solution @xmath150 obtained from the proposed algorithm , the central processor should generally perform joint compression in order to obtain the signals @xmath110 to be transmitted by the bss . however , as seen in sec .
[ sub : successive - compression ] , if the solution is such that the corner point conditions ( [ eq : corner points ] ) are satisfied for a given permutation @xmath161 of the bss indices , then the simpler successive estimation - compression structure of fig .
[ fig : successive comp ] can be leveraged instead .
we recall , from lemma [ lem : contra polymatroid ] , that in order to check whether the conditions ( [ eq : corner points ] ) are satisfied for some order @xmath161 , it is sufficient to observe which inequalities ( [ eq : original problem backhaul ] ) are satisfied with equality : if these inequalities correspond to the subsets @xmath224 for a given permutation @xmath161 , then the given solution corresponds to the corner point ( [ eq : corner points ] ) with the given @xmath161 . in our extensive numerical results
, we have consistently found this condition to be verified . as a result , in practice
, one can implement the compression strategy characterized by the calculated covariance @xmath117 by employing the implementation of fig .
[ fig : successive comp ] with the obtained ordering @xmath161 .
for reference , it is useful to consider the weighted sum - rate maximization problem with independent quantization noises as in @xcite-@xcite .
this is formulated as ( [ eq : original problem ] ) with the additional constraints @xmath225 since the constraints ( [ eq : independent quantization ] ) are affine , the mm algorithm in table algorithm 1 is still applicable by simply setting to zero matrices @xmath124 for @xmath125 as per ( [ eq : independent quantization ] ) .
so far , we have assumed that the central encoder has information about the global channel matrices @xmath63 for @xmath64 . in this subsection
, we discuss the robust design of the precoding matrix @xmath97 and the compression covariance @xmath117 in the presence of uncertainty at the central encoder regarding the channel matrices @xmath63 for @xmath64 . specifically , we focus on deterministic worst - case optimization under two different uncertainty models , namely singular value uncertainty @xcite and ellipsoidal uncertainty models ( see @xcite ( * ? ? ?
4.1 ) and references therein ) . while the singular value uncertainty model can be related via appropriate bounds to any normed uncertainty on the channel matrices , as discussed in (
v ) , the ellipsoidal uncertainty model is more accurate when knowledge of the covariance matrix of the csi error , due , e.g. , to estimation , is available ( * ? ? ?
4.1 ) . in the following ,
we briefly discuss both models . considering multiplicative uncertainty model of (
ii - a ) , the actual channel matrix @xmath63 toward each ms @xmath32 is modeled as @xmath226 where the matrix @xmath227 is the csi known at the central encoder and the matrix @xmath228 accounts for the multiplicative uncertainty matrix .
the latter is bounded as @xmath229 where @xmath230 is the largest singular value of matrix @xmath17 .
then , the problem of interest is to maximizing the worst - case weighted sum - rate over all possible uncertainty matrices @xmath231 for @xmath64 subject to the backhaul capacity ( [ eq : original problem backhaul ] ) and power constraints ( [ eq : original problem power ] ) , namely [ eq : worst - case problem ] @xmath232 the following lemma offers an equivalent formulation for problem ( [ eq : worst - case problem ] ) .
[ lem : problem singular uncertainty ] the problem ( [ eq : worst - case problem ] ) is equivalent to the problem ( [ eq : original problem ] ) with the channel matrix @xmath63 replaced with @xmath233 for @xmath64 .
we first observe that the uncertainty matrix @xmath231 affects only the corresponding rate function @xmath234 in ( [ eq : rate ms k computed ] ) .
therefore , the minimization versus matrices @xmath231 for @xmath64 in ( [ eq : worst - case problem objective ] ) can be performed separately for each @xmath32 by solving the problem @xmath235 . it can be now easily seen , following ( * ? ? ?
* theorem 1 ) , that the result of this minimization is obtained when @xmath231 is such that @xmath236 .
this concludes the proof .
based on lemma [ lem : problem singular uncertainty ] , one can hence solve problem ( [ eq : worst - case problem ] ) by using the mm algorithm in table algorithm 1 with only change of the channel matrices from @xmath237 to @xmath238 .
we now consider the ellipsoidal uncertainty model . to this end , for simplicity and following related literature ( see , e.g. , ( * ? ? ?
4.1 ) ) , we focus on multiple - input single - output ( miso ) case where each ms is equipped with a single antenna , i.e. , @xmath239 for @xmath64 .
thus , we denote the channel vector corresponding to each ms @xmath32 by @xmath240 .
the actual channel @xmath241 is then modeled as @xmath242 with @xmath243 and @xmath244 being the presumed csi available at the central encoder and the csi error , respectively .
the error vector @xmath244 is assumed to be bounded within the ellipsoidal region described as @xmath245 for @xmath64 with the matrix @xmath246 specifying the size and shape of the ellipsoid @xcite .
following the standard formulation , we consider here the `` dual '' problem of power minimization under signal - to - interference - plus - noise ratio ( sinr ) constraints for all mss ( see ( * ? ? ?
4.1 ) and references therein ) .
this problem is stated as [ eq : problem pmin ] @xmath247 where the coefficients @xmath248 are arbitrary weights , @xmath249 is the sinr constraint for ms @xmath32 , and we recall that we have @xmath200 for @xmath64 .
the problem ( [ eq : problem pmin ] ) is challenging since it contains infinite number of constraints in ( [ eq : problem pmin sinr constraint ] ) .
but , following the conventional _ s - procedure _
* appendix b.2 ) , we can translate the constraints ( [ eq : problem pmin sinr constraint ] ) into a finite number of linear constraints by introducing auxiliary variables @xmath250 for @xmath64 , as discussed in the following lemma .
[ lem : s lemma ] the constraints ( [ eq : problem pmin sinr constraint ] ) hold if and only if there exist constants @xmath251 such that the condition @xmath252-\beta_{k}\left[\begin{array}{cc } \mathbf{c}_{k } & \mathbf{0}\\ \mathbf{0 } & -1 \end{array}\right]\succeq\mathbf{0}\label{eq : s lemma constraint}\ ] ] is satisfied for all @xmath64 where we have defined @xmath253 for @xmath64
. it directly follows by applying the s - procedure ( * ? ? ?
* appendix b.2 ) . by transforming the constraint ( [ eq : problem pmin sinr constraint ] ) into ( [ eq
: s lemma constraint ] ) , we obtain a problem that falls again in the class of dc problems @xcite .
therefore , one can easily derive the mm algorithm , similar to table algorithm 1 , by linearizing the non - convex terms in the constraint ( [ eq : problem pmin backhaul constraints ] ) .
the algorithm is guaranteed to converge to a stationary point of problem ( [ eq : problem pmin ] ) ( see , e.g. , ( * ? ? ?
* theorem 1 ) , ( * ? ? ?
1.3.3 ) and ( * ? ?
* theorem 3 ) ) .
in this section , we discuss a simpler approach in which the precoding matrix @xmath97 is fixed a priori to some standard scheme , such as zf , mmse or weighted sum - rate maximizing precoding , by neglecting the compression noise .
the compression covariance @xmath117 is then designed separately so as to maximize the weighted sum - rate .
the precoding matrix @xmath97 is first selected according to some standard criterion @xcite-@xcite by neglecting the compression noise . a subtle issue arises when selecting the precoding matrix @xmath97 that requires some discussion .
specifically , the design of @xmath97 should be done by assuming a reduced power constraint , say @xmath254 for some @xmath255 for @xmath109 .
the power offset factor @xmath255 is necessary since the final signal @xmath110 transmitted by each bs @xmath38 is given by ( [ eq : gaussian test channel each bs ] ) and is thus the sum of the precoded signal @xmath256 and the compression noise @xmath112 .
therefore , if the power of the precoded part @xmath256 is selected to be equal to the power constraint @xmath257 , the compression noise power would be forced to be zero .
but this is possible only when the backhaul capacity grows to infinity due to ( [ eq : original problem backhaul ] ) . as a result , in order to make the compression feasible , one needs to properly select the parameters @xmath258 depending on the backhaul constraints .
having fixed the precoding matrix @xmath97 , the problem then reduces to solving problem ( [ eq : original problem ] ) only with respect to the compression covariance @xmath117 .
the obtained problem is thus a dc problem which can be tackled via the mm algorithm described in table algorithm 1 by limiting the optimization at step 2 only to matrix @xmath117 .
it is observed that , as discussed above , this problem may not be feasible if the parameters @xmath259 , @xmath109 , are too large . in practice
, one can set these parameters using various search strategies such as bisection .
in this section , we present numerical results in order to investigate the advantage of the proposed approach based on multivariate compression and on the joint design of precoding and compression as compared to the conventional approaches based on independent compression across the bss and separate design .
we will focus on the sum - rate performance @xmath260 ( i.e. , we set @xmath261 in ( [ eq : original problem objective ] ) ) .
we also assume that there is one ms active in each cell and we consider three cells , so that we have @xmath262 .
every bs is subject to the same power constraint @xmath263 and has the same backhaul capacity @xmath264 , i.e. , @xmath265 and @xmath266 for @xmath109 .
per - cell sum - rate versus the backhaul capacity @xmath264 for the circular wyner model @xcite with @xmath267 db and @xmath268.,width=453,height=340 ] we start by considering as a benchmark the performance in a simple circulant wyner model . in this model ,
all mss and bss have a single antenna and the channel matrices @xmath269 reduce to deterministic scalars given as @xmath270 for @xmath271 and @xmath272 $ ] for @xmath273 @xcite . in fig .
[ fig : graph wyner ] , we compare the proposed scheme with joint design of precoding and compression with state - of - the - art techniques , namely the compressed dpc of @xcite , which corresponds to using dpc precoding with independent quantization , and reverse compute - and - forward ( rcof ) @xcite . we also show the performance with linear precoding for reference .
it is observed that multivariate compression significantly outperforms the conventional independent compression strategy for both linear and dpc precoding .
moreover , rcof in @xcite remains the most effective approach in the regime of moderate backhaul @xmath264 , although multivariate compression allows to compensate for most of the rate loss of standard dpc precoding in the low - backhaul regime is due to the integer constraints imposed on the function of the messages to be computed by the mss @xcite .
] . in this subsection
, we evaluate the average sum - rate performance as obtained by averaging the sum - rate @xmath274 over the the realization of the fading channel matrices .
the elements of the channel matrix @xmath275 between the ms in the @xmath32th cell and the bs in the @xmath38th cell are assumed to be i.i.d .
complex gaussian random variables with @xmath276 in which we call @xmath277 the inter - cell channel gain .
moreover , each bs is assumed to use two transmit antennas while each ms is equipped with a single receive antenna . in the separate design
, we assume that the precoding matrix @xmath97 is obtained via the sum - rate maximization scheme in @xcite under the power constraint @xmath278 for each bs with @xmath279 selected as discussed in sec .
[ sub : separate precoding ] .
note that the algorithm of @xcite finds a stationary point for the sum - rate maximization problem using the mm approach , similar to table algorithm 1 without consideration of the quantization noises .
average sum - rate versus the power offset factor @xmath280 for the separate design of linear precoding and compression in sec .
[ sec : separate ] with @xmath281 db and @xmath282 db.,width=453,height=340 ] fig .
[ fig : graph gamma ] demonstrates the impact of the power offset factor @xmath280 on the separate design of linear precoding and compression described in sec .
[ sec : separate ] with @xmath281 db and @xmath282 db .
increasing @xmath280 means that more power is available at each bs , which generally results in a better sum - rate performance .
however , if @xmath280 exceeds some threshold value , the sum - rate is significantly degraded since the problem of optimizing the compression covariance @xmath117 given the precoder @xmath97 is more likely to be infeasible as argued in sec .
[ sub : separate precoding ] .
this threshold value grows with the backhaul capacity , since a larger backhaul capacity allows for a smaller power of the quantization noises . throughout the rest of this section ,
the power offset factor @xmath280 is optimized via numerical search .
average sum - rate versus the transmit power @xmath263 for linear precoding with @xmath283 bit / c.u . and @xmath282 db.,width=453,height=340 ] in fig .
[ fig : graph snr ] , the average sum - rate performance of the linear precoding and compression schemes is plotted versus the transmit power @xmath263 with @xmath283 bit / c.u . and
@xmath282 db .
it is seen that the gain of multivariate compression is more pronounced when each bs uses a larger transmit power .
this implies that , as the received snr increases , more efficient compression strategies are called for . in a similar vein ,
the importance of the joint design of precoding and compression is more significant when the transmit power is larger .
moreover , it is seen that multivariate compression is effective in partly compensating for the suboptimality of the separate design . for reference
, we also plot the cutset bound which is obtained as @xmath284 where @xmath285 is the sum - capacity achievable when the bss can fully cooperate under per - bs power constraints .
we have obtained the rate @xmath285 by using the inner - outer iteration algorithm proposed in ( * ? ? ?
it is worth noting that only the proposed joint design with multivariate compression approaches the cutset bound as the transmit power increases .
average sum - rate versus the transmit power @xmath263 for the joint design in sec .
[ sec : joint ] with @xmath283 bit / c.u . and @xmath282 db.,width=453,height=340 ] in fig .
[ fig : graph dpc ] , we compare two precoding methods , dpc and linear precoding , by plotting the average sum - rate versus the transmit power @xmath263 for the joint design in sec . [
sec : joint ] with @xmath283 bit / c.u . and @xmath282 db . for dpc , we have applied algorithm 1 with a proper modification for all permutations @xmath286 of mss indices @xmath127 and took the largest sum - rate . unlike the conventional broadcast channels with perfect backhaul links where there exists constant sum - rate gap between dpc and linear precoding at high snr ( see ,
e.g. , @xcite ) , fig .
[ fig : graph dpc ] shows that dpc is advantageous only in the regime of intermediate @xmath263 due to the limited - capacity backhaul links .
this implies that the overall performance is determined by the compression strategy rather than precoding method when the backhaul capacity is limited at high snr .
average sum - rate versus the backhaul capacity @xmath264 for linear precoding with @xmath281 db and @xmath282 db.,width=453,height=340 ] fig .
[ fig : graph ci ] plots the average sum - rate versus the backhaul capacity @xmath264 for linear precoding with @xmath281 db and @xmath282 db . it is observed that when the backhaul links have enough capacity , the benefits of multivariate compression or joint design of precoding and compression become negligible since the overall performance becomes limited by the sum - capacity achievable when the bss are able to fully cooperate with each other .
it is also notable that the separate design with multivariate compression outperforms the joint design with independent quantization for backhaul capacities larger than @xmath287 bit / c.u .
average sum - rate versus the inter - cell channel gain @xmath277 for linear precoding with @xmath283 bit / c.u . and @xmath281 db.,width=453,height=340 ]
finally , we plot the sum - rate versus the inter - cell channel gain @xmath277 for linear precoding with @xmath283 bit / c.u . and @xmath281 db in fig .
[ fig : graph alpha ] .
we note that the multi - cell system under consideration approaches the system consisting of @xmath29 parallel single - cell networks as the inter - cell channel gain @xmath277 decreases .
thus , the advantage of multivariate compression is not significant for small values of @xmath277 , since introducing correlation of the quantization noises across bss is helpful only when each ms suffers from a superposition of quantization noises emitted from multiple bss .
in this work , we have studied the design of joint precoding and compression strategies for the downlink of cloud radio access networks where the bss are connected to the central encoder via finite - capacity backhaul links . unlike the conventional approaches where
the signals corresponding to different bss are compressed independently , we have proposed to exploit multivariate compression of the signals of different bss in order to control the effect of the additive quantization noises at the mss .
the problem of maximizing the weighted sum - rate subject to power and backhaul constraints was formulated , and an iterative mm algorithm was proposed that achieves a stationary point of the problem .
moreover , we have proposed a novel way of implementing multivariate compression that does not require joint compression of all the bss signals but is based on successive per - bs estimation - compression steps .
robust design with imperfect csi was also discussed . via numerical results , it was confirmed that the proposed approach based on multivariate compression and on joint precoding and compression strategy outperforms the conventional approaches based on independent compression and separate design of precoding and compression strategies .
this is especially true when the transmit power or the inter - cell channel gain are large , and when the limitation imposed by the finite - capacity backhaul links is significant .
15 j. andrews , `` the seven ways hetnets are a paradigm shift , '' to appear in _ ieee comm . mag .
_ , mar . 2013 .
intel cor .
, `` intel heterogeneous network solution brief , '' solution brief , intel core processor , telecommunications industry .
j. segel and m. weldon , `` lightradio portfolio - technical overview , '' technology white paper 1 , alcatel - lucent .
s. liu , j. wu , c. h. koh and v. k. n. lau , `` a 25 gb / s(/km@xmath288 ) urban wireless network beyond imt - advanced , '' _ ieee comm . mag .
122 - 129 , feb . 2011 .
china mobile , `` c - ran : the road towards green ran , '' white paper , ver .
2.5 , china mobile research institute , oct . 2011 .
t. flanagan , `` creating cloud base stations with ti s keystone multicore architecture , '' white paper , texas inst .
, oct . 2011 .
ericsson , `` heterogeneous networks , '' ericsson white paper , feb . 2012 .
p. marsch , b. raaf , a. szufarska , p. mogensen , h. guan , m. farber , s. redana , k. pedersen and t. kolding , `` future mobile communication networks : challenges in the design and operation , '' _ ieee veh .
tech . mag .
_ , vol . 7 , no
16 - 23 , mar .
v. chandrasekhar , j. g. andrews and a. gatherer , `` femtocell networks : a survey , '' _ ieee comm . mag .
59 - 67 , sep . 2008 .
i. maric , b. bostjancic and a. goldsmith , `` resource allocation for constrained backhaul in picocell networks , '' in _ proc .
ita 11 _ , ucsd , feb .
s. h. lim , y .- h .
kim , a. e. gamal and s .- y .
chung , `` noisy network coding , '' _ ieee trans .
inf . theory _
3132 - 3152 , may 2011 .
a. e. gamal and y .- h .
network information theory _ , cambridge university press , 2011 .
a. sanderovich , o. somekh , h. v. poor and s. shamai ( shitz ) , `` uplink macro diversity of limited backhaul cellular network , '' _ ieee trans .
inf . theory _ ,
55 , no . 8 , pp . 3457 - 3478 , aug .
a. sanderovich , s. shamai ( shitz ) and y. steinberg , `` distributed mimo receiver - achievable rates and upper bounds , '' _ ieee trans .
inf . theory _ ,
55 , no . 10 , pp .
4419 - 4438 , oct .
a. del coso and s. simoens , `` distributed compression for mimo coordinated networks with a backhaul constraint , '' _ ieee trans .
wireless comm .
_ , vol . 8 , no .
9 , pp . 4698 - 4709 , sep .
l. zhou and w. yu , `` uplink multicell processing with limited backhaul via successive interference cancellation , '' in _ proc .
ieee glob .
( globecom 2012 ) _ , anaheim , ca , dec . 2012 . s .- h . park , o. simeone , o. sahin and s. shamai ( shitz ) , `` robust and efficient distributed compression for cloud radio access networks , '' _ ieee trans
692 - 703 , feb . 2013 .
m. grieger , s. boob and g. fettweis , `` large scale field trial results on frequency domain compression for uplink joint detection , '' in _ proc .
ieee glob .
( globecom 2012 ) _ , anaheim , ca , dec .
b. nazer , a. sanderovich , m. gastpar and s. shamai ( shitz ) , `` structured superposition for backhaul constrained cellular uplink , '' in _ proc . ieee intern .
inf . theory ( isit 2009 ) _ , seoul , korea , jun .
hong and g. caire , `` quantized compute and forward : a low - complexity architecture for distributed antenna systems , '' in _ proc .
theory workshop ( itw 2011 ) _ , paraty , brazil , oct .
o. simeone , o. somekh , h. v. poor and s. shamai ( shitz ) , `` downlink multicell processing with limited - backhaul capacity , '' _
eurasip j. adv .
proc . _ , 2009 .
p. marsch and g. fettweis , `` on downlink network mimo under a constrained backhaul and imperfect channel knowledge , '' in _ proc .
ieee glob .
( globecom 2009 ) _ , honolulu , hawaii , nov . 2009 .
m. h. m. costa , `` writing on dirty paper , '' _ ieee trans .
inf . theory _ ,
439 - 441 , may 1983 .
hong and g. caire , `` reverse compute and forward : a low - complexity architecture for downlink distributed antenna systems , '' in _ proc .
ieee intern .
inf . theory ( isit 2012 ) _ , cambridge , ma , jul .
b. l. ng , j. s. evans , s. v. hanly and d. aktas , `` distributed downlink beamforming with cooperative base stations , '' _ ieee trans .
inf . theory _ , vol .
5491 - 5499 , dec . 2008
. i. sohn , s. h. lee and j. g. andrews , `` belief propagation for distributed downlink beamforming in cooperative mimo cellular networks , '' _ ieee trans .
wireless comm .
4140 - 4149 , dec . 2011 .
r. zakhour and d. gesbert , `` optimized data sharing in multicell mimo with finite backhaul capacity , '' _ ieee trans
proc . _ , vol .
6102 - 6111 , dec .
o. simeone , n. levy , a. sanderovich , o. somekh , b. m. zaidel , h. v. poor and s. shamai ( shitz ) , `` cooperative wireless cellular systems : an information - theoretic view , '' _ foundations and trends in communications and information theory _ , vol
. 8 , nos . 1 - 2 , pp . 1 - 177 , 2012 .
l. zhang , r. zhang , y .- c .
liang , y. xin and h. v. poor , `` on gaussian mimo bc - mac duality with multiple transmit covariance constraints , '' _ ieee trans .
inf . theory _ ,
2064 - 2078 , apr . 2012 .
x. zhang , j. chen , s. b. wicker and t. berger , `` successive coding in multiuser information theory , '' _ ieee trans .
inf . theory _ ,
53 , no . 6 , pp . 2246 - 2254 , jun .
r. zhang , `` cooperative multi - cell block diagonalization with per - base - station power constraints , '' _
ieee j. sel .
areas comm .
1435 - 1445 , dec .
2010 . m. hong , r .- y .
sun , h. baligh and z .- q .
luo , `` joint base station clustering and beamformer design for partial coordinated transmission in heterogeneous networks , '' _ ieee j. sel .
areas comm .
226 - 240 , feb . 2013 . c. t. k. ng and h. huang ,
`` linear precoding in cooperative mimo cellular networks with limited coordination clusters , '' _ ieee j. sel .
areas comm .
9 , pp . 1446 - 1454 , dec .
m. hong , q. li , y .- f .
liu and z .- q .
luo , `` decomposition by successive convex approximation : a unifying approach for linear transceiver design in interfering heterogeneous networks , '' arxiv:1210.1507 .
u. erez and s. t. brink , `` a close - to - capacity dirty paper coding scheme , '' _ ieee trans .
inf . theory _ ,
3417 - 3432 , oct .
g. d. forney , jr .
, `` shannon meets wiener ii : on mmse estimation in successive decoding schemes , '' arxiv:0409011v2 .
d. n. c. tse and s. v. hanly , `` multiaccess fading channels - part i : polymatroid structure , optimal resource allocation and throughput capacities , '' _ ieee trans .
inf . theory _ ,
44 , no . 7 , pp .
2796 - 2815 , nov .
a. beck and m. teboulle , `` gradient - based algorithms with applications to signal recovery problems , '' in _
convex optimization in signal processing and communications _ ,
y. eldar and d. palomar , eds .
42 - 88 , cambridge university press . 2010 .
g. scutari , f. facchinei , p. song , d. p. palomar and j .- s .
pang , `` decomposition by partial linearization : parallel optimization of multi - agent systems , '' arxiv:1302.0756v1 .
s. loyka and c. d. charalambous , `` on the compound capacity of a class of mimo channels subject to normed uncertainty , '' _ ieee trans .
inf . theory _ , vol .
2048 - 2063 , apr .
2012 . c. shen , t .- h .
chang , k .- y .
wang , z. qiu and c .- y .
chi , `` distributed robust multicell coordinated beamforming with imperfect csi : an admm approach , '' _ ieee trans .
sig . proc .
60 , no . 6 , pp .
2988 - 3003 , jun . 2012 .
e. bjornson and e. jorswieck , `` optimal resource allocation in coordinated multi - cell systems , '' _ foundations and trends in communications and information theory _ ,
vol . 9 , nos . 2 - 3 , pp .
113 - 381 , 2013 .
s. boyd and l. vandenberghe , _ convex optimization _ , cambridge university press , 2004 .
d. gesbert , s. hanly , h. huang , s. shamai ( shitz ) , o. simeone and w. yu , `` multi - cell mimo cooperative networks : a new look at interference , '' _ ieee j. sel .
areas comm .
1380 - 1408 , dec . 2010
h. huh , h. c. papadopoulos and g. caire , `` multiuser miso transmitter optimization for intercell interference mitigation , '' _ ieee trans .
4272 - 4285 , aug .
j. lee and n. jindal , `` high snr analysis for mimo broadcast channels : dirty paper coding versus linear precoding , '' _ ieee trans .
inf . theory _ ,
4787 - 4792 , dec . 2007 | this work studies the joint design of precoding and backhaul compression strategies for the downlink of cloud radio access networks . in these systems
, a central encoder is connected to multiple multi - antenna base stations ( bss ) via finite - capacity backhaul links . at the central encoder ,
precoding is followed by compression in order to produce the rate - limited bit streams delivered to each bs over the corresponding backhaul link . in current state - of - the - art approaches ,
the signals intended for different bss are compressed independently . in contrast
, this work proposes to leverage joint compression , also referred to as multivariate compression , of the signals of different bss in order to better control the effect of the additive quantization noises at the mobile stations ( mss ) .
the problem of maximizing the weighted sum - rate with respect to both the precoding matrix and the joint correlation matrix of the quantization noises is formulated subject to power and backhaul capacity constraints .
an iterative algorithm is proposed that achieves a stationary point of the problem . moreover , in order to enable the practical implementation of multivariate compression across bss , a novel architecture is proposed based on successive steps of minimum mean - squared error ( mmse ) estimation and per - bs compression .
robust design with respect to imperfect channel state information is also discussed . from numerical results , it is confirmed that the proposed joint precoding and compression strategy outperforms conventional approaches based on the separate design of precoding and compression or independent compression across the bss .
cloud radio access network , constrained backhaul , precoding , multivariate compression , network mimo . |
y. n. thanks the frankfurt institute of advanced studies where part of this work was done .
this work was supported in part by the grants - in - aid for scientific research from jsps ( nos .
15k05079and 15k05098 ) , h. n. has received funding from the european union s horizon 2020 research and innovation programme under the marie sklodowska - curie grant agreement no .
655285 and from the helmholtz international center for fair within the framework of the loewe program launched by the state of hesse .
computational resources have been provided by the center for scientific computing ( csc ) at the goethe - university of frankfurt and gsi ( darmstadt ) .
d. h. rischke , prog . part .
phys . * 52 * , 197 ( 2004 ) ; m. a. stephanov , prog .
suppl . * 153 * , 139 ( 2004 ) [ int .
j. mod .
a * 20 * , 4387 ( 2005 ) ] ; k. fukushima and t. hatsuda , rept .
phys . * 74 * , 014001 ( 2011 ) ; k. fukushima and c. sasaki , prog . part .
phys . * 72 * , 99 ( 2013 ) .
j. brachmann , s. soff , a. dumitru , h. stoecker , j. a. maruhn , w. greiner , l. v. bravina and d. h. rischke , phys .
c * 61 * , 024909 ( 2000 ) .
j. brachmann , a. dumitru , h. stoecker and w. greiner , eur .
j. a * 8 * , 549 ( 2000 ) . y. b. ivanov , e. g. nikonov , w. noerenberg , a. a. shanenko and v. d. toneev , heavy ion phys .
* 15 * , 117 ( 2002 ) ; v. d. toneev , e. g. nikonov , b. friman , w. norenberg and k. redlich , eur .
j. c * 32 * , 399 ( 2003 ) .
l. p. csernai and d. rohrich , phys .
b * 458 * , 454 ( 1999 ) .
l. p. csernai , a. anderlik , c. anderlik , v. k. magas , e. molnar , a. nyiri , d. rohrich and k. tamosiunas , acta phys . hung . a * 22 * , 181 ( 2005 ) . l. adamczyk _ et al . _
[ star collaboration ] , phys .
lett . *
112 * , no .
16 , 162301 ( 2014 ) .
p. shanmuganathan [ star collaboration ] , arxiv:1512.09009 [ nucl - ex ] .
s. singha , p. shanmuganathan and d. keane , arxiv:1610.00646 [ nucl - ex ] .
g. f. bertsch , h. kruse and s. d. gupta , phys .
c * 29 * , 673 ( 1984 ) erratum : [ phys .
c * 33 * , 1107 ( 1986 ) ] ; g. f. bertsch and s. das gupta , phys .
rept . * 160 * , 189 ( 1988 ) ; h. kruse , b. v. jacak , j. j. molitoris , g. d. westfall and h. stoecker , phys .
c * 31 * , 1770 ( 1985 ) ; h. kruse , b. v. jacak and h. stoecker , phys . rev .
* 54 * , 289 ( 1985 ) ; w. cassing , v. metag , u. mosel and k. niita , phys . rept . * 188 * , 363 ( 1990 ) . c. m. ko , q. li and r. c. wang , phys .
lett . * 59 * , 1084 ( 1987 ) ; h. t. elze , m. gyulassy , d. vasak , h. heinz , h. stoecker and w. greiner , mod .
lett . a * 2 * , 451 ( 1987 ) ; b. blaettel , v. koch and u. mosel , rept .
prog . phys .
* 56 * , 1 ( 1993 ) .
m. gyulassy , k. a. frankel and h. stoecker , phys .
b * 110 * , 185 ( 1982 ) .
d. e. kahana , d. keane , y. pang , t. schlagel and s. wang , phys .
lett . * 74 * , 4404 ( 1995 ) ; d. e. kahana , y. pang and e. v. shuryak , phys .
c * 56 * , 481 ( 1997 ) .
b. a. li and c. m. ko , phys .
c * 58 * , 1382 ( 1998 ) .
p. danielewicz , p. b. gossiaux and r. a. lacey , in `` les houches 1998 , nuclear matter in different phases and transitions '' 69 - 84 [ ucl - th/9808013 ] ; p. danielewicz , in `` hirschegg 1999 , multifragmentation '' 263 - 272 [ nucl - th/9902043 ] .
y. nara , n. otuka , a. ohnishi , k. niita and s. chiba , phys .
c * 61 * , 024901 ( 2000 ) .
h. sorge , phys .
c * 52 * , 3291 ( 1995 ) .
s. a. bass _ et al .
. part .
phys . *
41 * , 255 ( 1998 ) .
m. bleicher _ et al .
_ , j. phys .
g * 25 * , 1859 ( 1999 ) .
p. f. kolb , j. sollfrank and u. w. heinz , phys .
b * 459 * , 667 ( 1999 ) ; p. f. kolb , j. sollfrank and u. w. heinz , phys .
c * 62 * , 054909 ( 2000 ) ; h. song and u. w. heinz , phys .
c * 77 * , 064901 ( 2008 ) .
j. steinheimer , s. schramm and h. stoecker , j. phys .
g * 38 * , 035001 ( 2011 ) ; j. steinheimer , s. schramm and h. stoecker , phys .
c * 84 * , 045208 ( 2011 ) ; p. rau , j. steinheimer , s. schramm and h. stoecker , phys .
c * 85 * , 025204 ( 2012 ) .
h. liu _ et al .
_ [ e895 collaboration ] , phys . rev .
lett . * 84 * , 5488 ( 2000 ) ; h. appelshauser _ et al . _
[ na49 collaboration ] , phys .
. lett . * 80 * , 4136 ( 1998 ) ; c. alt _ et al . _
[ na49 collaboration ] , phys .
c * 68 * , 034903 ( 2003 ) . | we study the sensitivities of the directed flow in au+au collisions on the equation of state ( eos ) , employing the transport theoretical model jam .
the eos is modified by introducing a new collision term in order to control the pressure of a system by appropriately selecting an azimuthal angle in two - body collisions according to a given eos .
it is shown that this approach is an efficient method to modify the eos in a transport model .
the beam energy dependence of the directed flow of protons is examined with two different eos , a first - order phase transition and crossover .
it is found that our approach yields quite similar results as hydrodynamical predictions on the beam energy dependence of the directed flow ; transport theory predicts a minimum in the excitation function of the slope of proton directed flow and does indeed yield negative directed flow , if the eos with a first - order phase transition is employed .
our result strongly suggests that the highest sensitivity for the critical point can be seen in the beam energy range of @xmath0 gev .
one of the most challenging problems in high energy heavy ion collisions is to map out the qcd phase diagram from low to high baryon densities . in particular ,
the main interest is to discover a first and/or second order phase transition together with the critical point of qcd matter at finite baryon chemical potentials , which was predicted by several effective models based on qcd @xcite . in order to explore qcd matter in the full range of temperatures and baryon densities ,
it is necessary to measure various observables for a wide range of beam energies .
a non - trivial beam energy dependence of various observables such as the @xmath1 ratio @xcite , higher order of the net - proton number cumulants @xcite and the slope of the directed flow have been reported from the na49 collaboration and the star beam energy scan ( bes ) program @xcite .
we note that such a beam energy dependence usually is not explained by the standard hadronic transport models
. there have been attempts to take into account a change of degrees of freedom ( into quarks and gluons ) @xcite in transport models , or transport + hydrodynamics hybrid models @xcite .
for instance , hybrid models usually show an improved description of strange particle ratios and collective flow @xcite the collective transverse flow developed in the heavy ion collisions is considered to be sensitive to the equation of state ( eos ) of qcd matter , since it reflects the pressure gradients in the early stages of the reaction .
especially , the softest region in the eos with a ( first - order ) phase transition is expected to have significant influence on the directed flow @xmath2 of nucleons @xcite .
hydrodynamic predictions have shown that the excitation function of directed flow exhibits a local minimum at a specific beam energy characteristic for the first - order phase transition @xcite .
furthermore , the slope of the directed flow of nucleons becomes negative at the softest point of the eos .
@xcite .
thus the collapse of directed flow may be a signature of the qcd first - order phase transition at high baryon densities .
indeed , such a behavior has been observed in the bes program by the star collaboration @xcite but no microscopic hadronic transport model is yet able to quantitatively explain the negative slope of protons at @xmath3 and 19.6 gev @xcite . only at higher beam energies ,
the microscopic transport models rqmd @xcite , urqmd @xcite , and phsd / hsd @xcite show a negative slope of proton directed flow without a phase transition .
we would like to note that studies including a hadronic mean field interaction also have not been able to describe the negative slope of the proton directed flow @xcite .
the excitation function of the directed flow has also been examined by other approaches such as the urqmd hybrid model @xcite , three fluid models @xcite , and jam with an attractive scattering style @xcite , but none of the models is able to give a quantitative satisfactory description of the star data .
therefore it is still premature to make unambiguous interpretations of the star data , and more reliable models have to be developed in order to simulate the spacetime evolution of the system created in heavy ion collisions at intermediate beam energies and to understand the underlying collision dynamics .
there is a long history of implementing an eos into transport theoretical approaches .
the boltzmann - uehling - uhlenbeck ( buu ) model @xcite was developed first followed by quantum molecular dynamics ( qmd ) @xcite to test various eos in a transport model that includes nuclear mean field potentials .
later a relativistic extension of both buu and qmd was developed which are called rbuu @xcite and rqmd @xcite .
it is also known that the eos can be controlled by changing the scattering style in the two - body collision term @xcite .
for example , choosing a repulsive orbit in two - body collisions can simulate the effect of a repulsive potential .
later , effects of a phase transition have been investigated by including a mean field interaction @xcite or using attractive orbits in two - body collisions @xcite .
the theoretically favored method is to include mean fields to vary the eos , and in principle any mean field can be incorporated provided that the interaction lagrangian is known .
it is , however , practically very challenging to implement a mean field in a dynamical simulation , and one often relies on the simplified phenomenological form of potentials . on the other hand ,
an advantage of the second approach is that we can include any kind of eos computed by other methods such as effective models or lattice qcd . in this work
, we shall follow the second approach to examine effects of the eos on the directed flow within the microscopic transport approach jam by modifying the scattering style in the two - body collision term .
we do not include explicitly partonic degree of freedom , as for example in ampt @xcite or phsd @xcite , since the transition from one set of degrees of freedom is difficult and usually introduced new uncertainties in the model .
see also ref .
@xcite for a recent attempt to implement a first order phase transition into a parton cascade model . in jam
@xcite , particle production is modeled by the excitation and decay of resonances and strings similar to the other models @xcite .
secondary products from decays can interact with each other via binary collisions .
a detailed description of the jam model can be found in ref .
@xcite . in the standard cascade approach the two body collision term
is implemented so that it does not generate additional pressure .
namely , the azimuthal angle in the two - body collision is randomly chosen .
it is well known that the pressure can be controlled by changing the scattering style . here
we employ a method similar to that proposed in ref .
@xcite in which the momentum change in each two - body collision between particle @xmath4 and @xmath5 at the space - time coordinates of @xmath6 and @xmath7 is related to the pressure @xmath8 , where @xmath9 is the free streaming part of the local pressure . motivated by the virial theorem @xcite ,
the formula is given by @xmath10 where @xmath11 is the lorentz invariant local particle density obtained by @xmath12 , @xmath13 , @xmath14 is the proper time interval between successive collisions , and @xmath15 is the energy - momentum change of the particle @xmath4 .
@xmath9 can be computed from the energy - momentum tensor @xmath16 : @xmath17 , where @xmath18 with the projector of @xmath19 , where @xmath20 is a hydrodynamics velocity defined by the landau and lifshitz s definition . in this way
, we control the pressure such that @xmath21 coincides with a given eos at an energy density @xmath22 .
there are three ways to satisfy the constraint eq.([eq : pre ] ) to control the pressure in two - body collisions ; 1 .
change the scattering angle , 2 . change the azimuthal angle , 3 .
change the magnitude of the relative momentum by modifying the masses of the outgoing particles . in ref .
@xcite , an additional elastic scattering is generated after any of the standard two - body scatterings , and it was found that more elliptic flow is generated .
it is known that the shear viscosity is sensitive to the scattering angle @xcite .
thus , it is unclear whether the eos alone is responsible for the modification of elliptic flow , since transport coefficients are also modified at the same time by additional extra elastic scatterings .
we therefore avoid changing the scattering angle in this work in order to model the eos and effectively keeping the transport coefficients as unchanged as possible .
the idea of changing the outgoing mass is attractive in the sense that it may effectively simulate a dropping mass due to chiral symmetry restoration ( csr ) .
but it is not straightforward how to relate it to csr within our approach and hadrons would eventually have to retain their vacuum masses before escaping the dense system .
thus , in this work , we adapt the method of only changing the azimuthal angle in the two - body collision so that the constraint eq.([eq : pre ] ) is fulfilled .
we test two types of eos . for the first - order phase transition
, the hadron resonance gas phase contains all hadronic resonances up to 2 gev , and a baryon density @xmath23 dependent vector type repulsive mean field @xmath24 with @xmath25 gev @xmath26 is included as described in @xcite .
the qgp phase is modeled by the ideal gas of massless quarks and gluons with a bag constant of @xmath27 mev which leads to a phase transition temperature of @xmath28 mev at vanishing net baryon density . for the crossover eos
, we use the chiral model eos from ref .
@xcite in which eos at vanishing and finite baryon density is consistent with a smooth crossover transition , as found in lattice qcd results . and energy density @xmath29 in mid - central au+au collisions ( 10 - 40% ) at @xmath3 gev from jam cascade mode ( squares ) , jam with first - order eos ( triangles ) and with crossover eos ( circles ) .
initial points corresponds to the time 0.65 fm / c after touching of two nuclei , and time interval is 0.25 fm / c .
arrow indicates the direction of the reaction .
pressure and energy density are averaged over collision points in a cylindrical volume of transverse radius 3 fm and longitudinal distance of 3 fm centered at the origin .
, width=321 ] in fig .
[ fig : eos11 ] , the time evolution of the local isotropic pressure and energy density , averaged over events , in mid - central au+au collision at @xmath3 gev are compared for the different eos used .
the jam standard eos is the stiffest , which is due to the chemical non - equilibrium in the initial pre - equilibrium evolution as pointed out in ref .
@xcite .
a pronounced softening can be seen in the case of the first order phase transition at around the energy density of @xmath30 gev/@xmath26 as we expected . on the other hand ,
the effect of softening is weaker for the crossover eos . in the construction of the crossover eos from a chiral model , correct nuclear saturation properties at zero temperature where required , resulting in the stiff hadronic eos ,
compared to the standard jam eos which is expected to be close to an ideal hadron resonance gas eos .
this property leads effectively to the inclusion of repulsive potential in the hadronic phase in jam simulations .
we now compute the directed flow in mid - central au+au collisions , and compare the two eos described above . in the simulation , we choose the impact parameter range @xmath31 fm for mid - central collisions . in fig .
[ fig : v1_11 ] , we show the rapidity dependence of the proton ( upper panel ) and pion ( lower panel ) directed flows in mid - central au+au collision at @xmath3 gev .
the pion directed flow is not sensitive to the eos , all models are able reproduce the pion data , which means the pion directed flow is not useful to probe the early time dynamics of the collision . as we showed in ref .
@xcite , that the jam cascade calculation for protons shows a significantly larger flow than the star data .
it is seen that the proton directed flow from jam with both the first order phase transition and crossover eos are significantly reduced at mid - rapidity , and close to the data , indicating the importance of the eos effects in the early phase .
however , both eos still predict a positive @xmath32-slope contrary to the star data . for protons and pions in semi - central au+au collisions for @xmath3 gev from jam simulations with different eoss .
the solid line presents the jam with the eos with 1st order phase transition , and the dashed line presents the jam with the crossover eos .
the dotted line is for the standard jam result .
the star data are taken from ref .
@xcite .
, width=302 ] for protons in semi - central au+au collisions for @xmath3 gev from jam simulations by choosing attractive orbit with a probability given by eq .
( [ eq : attprob ] ) .
the solid line presents the jam with the eos with 1st order phase transition , and the dashed line presents the jam with the crossover eos .
upper panel : the results from baryon density independent eos used in ref .
@xcite .
lower panel : the results from baryon density dependent eos .
the star data are taken from ref .
@xcite . ,
width=302 ] we now check if the microscopic details are important for the generation of directed flow . in order to see systematics on the modeling in controlling a eos
, we show in fig .
[ fig : v1att11 ] , the rapidity dependence of proton directed flow at @xmath3 gev from the method of choosing attractive orbit in two - body scattering with the probability : @xmath33 as proposed in ref .
@xcite , where @xmath34 is a pressure as a function of energy density @xmath29 from a given eos . as we simply choose attractive orbit according to the probability given by eq .
( [ eq : attprob ] ) in this method , pressure at each two - body collision is not equal to the pressure of a given eos , and it fluctuates a lot .
only average pressure of the system coincides with the given eos . with this approach ,
one see that proton directed flow is less suppressed compared with our new method .
we also show the directed flow obtained by the baryon density independent eos as used in ref .
@xcite in the upper panel of fig .
[ fig : v1att11 ] .
baryon density dependent eos used in this work yields less suppression of the proton directed flow as we expected , since our eos predicts less softening at finite baryon densities . in this work
, we estimate the energy - momentum tensor by including constituent quarks .
technically it is done by counting leading hadrons that have original constitute quarks within a formation time multiplying the reduction factor of one - third ( two - third ) or one - half depending on the number of constituent quarks inside baryons or mesons . in the previous calculations in ref .
@xcite , we did not take into account the reduction factors in leading hadrons which leads to larger energy densities . to see the sensitivities of the different assumptions in defining the energy density within our model , the proton directed flow from the calculation in which only formed hadrons
are taken into account for the estimation of energy density @xcite is presented in fig .
[ fig : v1_11had ] . in this case
, energy density corresponds to hadronic energy density , and it is less than the default calculation , in which the contributions from constituent quarks are included , and as a result , we have more softening of eos and the slope of proton directed flow at @xmath3 gev becomes negative .
however , we do not see any sensitivities between the first order phase transition and the crossover eos at @xmath3 gev .
we also note that pion directed flow from jam deviates from the data when we use hadronic energy density . , but only formed hadrons are taken into account for the estimation of energy density .
, width=302 ] in the following we present the beam energy dependence of the slope of the directed flow . in fig .
[ fig : v1slope ] we show the slope of directed flow @xmath35 of protons in mid - central collisions from the standard jam cascade and jam with a first - order phase transition and with a crossover eos , in au+au collisions in comparison with the data from star , na49 and e895 collaborations @xcite .
the slope is obtained by fitting the rapidity dependence of @xmath32 to a cubic equation @xmath36 in the rapidity interval @xmath37 .
the standard jam cascade calculation predicts a minimum at around ags energies which was first reported in ref .
@xcite within the urqmd approach .
the decrease of the directed flow in the standard cascade approach can be understood by the rapid change in degree of freedoms due to the excitation of hadronic resonances up to 2 gev .
note that phsd does not have such minimum @xcite , since there are only a few number of hadronic resonances included in the phsd model .
it is important to notice that the slope of the proton directed flow obtained by the cascade model is still positive at the minimum point . at higher energies up to @xmath38 gev
, the jam standard model overestimates the slope of the proton directed flow as already reported in ref .
@xcite . in the case of the eos with the first order phase transition ( jam-1.opt )
, we also see the minimum in the excitation function of the proton directed flow at almost the same beam energy as the cascade model .
in addition , the slope is now negative as predicted by hydrodynamical approaches .
the beam energy dependence is very similar to the pure hydrodynamical simulation except that the magnitude of the minimum in jam-1.opt is about a factor 5 smaller than the ideal hydrodynamical prediction @xcite , which may be related to the finite viscosity in the transport approach .
the local minimum predicted by the jam-1.opt is located at a slightly higher beam energy than in the one - fluid model prediction ( @xmath39 gev ) in which the strong coupling of the fluids lead to an almost instantaneous full stopping and maximum energy deposition , different than in the three - fluid model in ref .
@xcite in which finite stopping power of nuclear matter is taken into account .
we note that the local minimum predicted by the three - fluid model in ref .
@xcite shows @xmath40 gev .
probably the location of minimum depends both on the eos and the degree of stopping and its modeling . in ref .
@xcite , the eos in the qgp phase is modeled by a quasi - particle approximation with mean field potential @xcite .
it is interesting to notice that the art buu approach with the first order phase transition also exhibits minimum with a negative slope @xcite which supports that our method effectively handles the effect of the eos . on the other hand , in the case of a crossover eos ( jam-@xmath41-over )
, there is still a local minimum at @xmath42 gev , but it is not so pronounced .
thus our approach supports the idea that a large negative slope of the proton directed flow would be a good observable to identify a strong softening in the early phase of the collision due to the first order phase transition . however , the minimum observed from our approach is located at a beam energy similar to that predicted by hydrodynamical simulations @xcite and much lower than the minimum measured by the star .
if the eos we employ is close to the true eos in nature , we do not see the connection between the softest point of the eos and the star data . in order to describe the minimum at a larger beam energy
one would need a very soft eos at very high baryon densities , much higher than what is found in the currently used eos , in order to shift the minimum to the higher beam energies . within our analysis ,
the reduction of the proton directed flow at @xmath43 gev is essentially related to the early pressure in the pre - equilibrium stages of the reaction .
the reason why jam-1.opt yields stronger flow than jam-@xmath41-over at @xmath43 gev is simply because of the fact that jam-1.opt in the current study assumes the massless ideal quark - gluon eos at the qgp phase , while crossover eos is consistent with the lattice qcd data at the vanishing baryon chemical potentials . even at the finite baryon chemical potentials , our crossover eos is softer than the massless ideal quark - gluon eos at high energy densities . since a massless ideal qgp eos
is not supported by the lattice qcd calculations , we need to test more realistic eos in order to see the difference between a first order phase transition and crossover transition in the future .
also it has been pointed out that the usual description of baryon stopping in string models in a transport model assumes essentially an instant deceleration of the leading baryons because strings are immediately decay into hadrons with a formation time . in @xcite
it was argued that a constant deceleration would lead to a very different initial density profile in coordinate space , i.e. a smaller density at mid - rapidity , which may transform into a different time evolution of the particle flow .
besides an uncertainty of the eos , there is a model uncertainty regarding the handling of static equilibrium eos @xmath34 within a non - equilibrium dynamical framework in which only effective local energy density can be obtained . to check how the directed flow is affected by the different definition of the energy density ,
we compute the slope of the proton directed flow by using hadronic energy density by counting only formed hadrons in the estimation of energy - momentum tensor which may be regarded as a possible lower limit of the energy density in the model .
the results from jam-1.opt and jam-@xmath41-over simulations are shown in the lower panel of fig .
[ fig : v1slope ] .
overall tendency does not change with this approach , but the slopes of proton directed flow at higher beam energies becomes negative compatible with the data up to @xmath44 gev except the point at @xmath45 gev .
we demonstrate that the directed flow of proton is highly sensitive to the eos at the beam energy range of @xmath46 gev , where no experimental data exists between 4.7 and 7.7 gev .
it is utmost important to measure the directed flow in this range in order to reveal the phase structure of qcd at the highest baryon densities . in summary
, we have investigated the eos dependence of the excitation function of the directed flow of protons within a microscopic transport approach by employing an energy density dependent eos modified collision term .
we showed that our approach provides an efficient method to control the eos in a microscopic transport model , and our result yields qualitatively similar result as the pure hydrodynamical predictions .
as predicted by hydrodynamical approach , the minimum of the directed flow , from a strong first order phase transition , is located at much lower beam energy than the star data . in order to distinguish between a first order phase transition and crossover transition in the directed flow data
, we need to examine different eos from models consistent with the lattice qcd up to finite baryon chemical potentials accessible in the current lattice qcd calculations . a systematic study including strange hadrons
should be addressed since strange hadrons follow a more complex pattern due to the different cross sections @xcite . in the future , it is interesting to analysis star data by explicitly implementing the mean field with phase transition into a transport model .
it is also interesting to investigate the effect of dropping of hadron mass due to csr on the directed flow .
we suggest to study the directed flow in the beam energy region of @xmath47 gev which may give suitable signatures to study the properties of eos at highest baryon density , where the softening effect could be best manifested .
future experiments such as the bes ii at rhic @xcite , fair @xcite , nica @xcite , and j - parc @xcite should provide important information on the properties of high dense matter created in the heavy ion collisions . |
in this paper , we study the dynamics of holomorphic endomorphisms of the complex projective space @xmath4 .
denote @xmath4 by @xmath5 from now on and let @xmath0 be an endomorphism of @xmath5 of algebraic degree @xmath6 .
such a map admits a unique invariant probability measure @xmath7 of maximal entropy @xmath8 , called the equilibrium measure . the most chaotic part of the dynamics is concentrated on @xmath9 .
we refer to @xcite for an introduction to this theme of research .
however , as opposed to dimension 1 , chaotic dynamics can also occur outside @xmath9 .
a basic non trivial dynamical phenomenon outside @xmath9 is that of attracting sets and attractors .
we refer to @xcite for some properties and examples of attracting sets and attractors for endomorphisms of @xmath5 and to @xcite for a basic structural description of the dynamics on the julia set .
see also @xcite for a detailed study of a class of algebraic attractors ( notice that the definition of an attractors is slightly more general there ) .
it is quite easy to find algebraic attracting set and attractors . on @xmath10 ,
all attractors are algebraic .
let @xmath11 be homogeneous polynomials of degree @xmath12 in @xmath13 with a single common zero @xmath14 , then the line at infinity @xmath15 ; t=0 \rbrace$ ] is an attracting set for @xmath16\mapsto [ p(z , w):q(z , w):t^d]$ ] and it is an attractor if the julia set of @xmath17\mapsto [ p(z , w):q(z , w)]$ ] is the whole riemann sphere , see @xcite . the first example of non algebraic attractor in @xmath3 ( resp .
@xmath5 , with @xmath18 ) was found by m. jonsson and b. weickert ( resp .
f. rong ) , see @xcite ( resp .
so far , most previously known examples of non algebraic attracting sets are , in a sense , of codimension 1 and occur for maps of the form @xmath19\mapsto [ p(z , w):q(z , w):t^d+\varepsilon r(z , w)],\ ] ] see @xcite .
a precise notion of the dimension of an attracting set can easily be formalized , see definition [ def dimension ] below .
dinh @xcite constructed , under some mild assumptions , a natural positive closed current supported on the attracting set .
this will be referred as _ the attracting current_. j. taflin @xcite , under additional hypotheses , proved that this current is the unique positive closed current supported on the attracting set .
it remains an interesting problem to understand the size " of the attracting set in a potential theoretic sense .
for instance , dinh @xcite asks whether the quasi - potential of the attracting current is always unbounded . for the basic example @xmath16\mapsto [ p(z , w):q(z , w):t^d]$ ] ,
the attracting set is the line @xmath20 thus it supports a unique positive closed current of bidegree ( 1,1 ) whose quasi - potential is unbounded . in this paper
, we address this problem , by introducing the concept of a mapping of small topological degree on an attracting set ( see definition [ def petit dt ] ) .
this notion was inspired by iteration theory of rational maps , see [ ddg1 - 3 ] .
the condition of being of small topological degree provides a conceptually simple framework to provide non algebraic attracting sets of any dimension ( see proposition [ prop non alg ] ) . in codimension 1
, this condition implies that the attracting set is non pluripolar . for this
we prove that it supports a positive closed current with bounded quasi - potential ( see theorem [ th potentiel bourne ] ) .
this answers dinh s question by the negative . on the other hand
, the condition of small topological degree is not sufficient in higher codimension to assure that the attracting set is non pluripolar ( see theorem [ th att cont in hyperplan ] ) .
the examples of small topological maps on attracting sets are abundant . more precisely we prove the following theorem : [ th intro ] denote by @xmath21 the quasi - projective variety of triples @xmath22 of homogeneous polynomials of degree @xmath23 in @xmath13 , such that @xmath14 is the single common zero of @xmath11 .
there exists a zariski open set @xmath24 such that if @xmath25 then for all @xmath26 small enough , @xmath27 , the endomorphism @xmath0 of @xmath3 defined by @xmath19\mapsto [ p(z , w):q(z , w):t^d + \varepsilon r(z , w)]\ ] ] admits an attracting set @xmath2 on which @xmath0 is of small topological degree . moreover ,
@xmath2 support a positive closed current @xmath28 of bidegree @xmath29 which admits a bounded quasi - potential .
of course , this theorem can not be true for every @xmath30 .
for example , @xmath31 \mapsto [ z^2 : w^2:t^2+\varepsilon z^2]$ ] is not of small topological degree on an attracting set , even for @xmath32 . indeed , the attracting set is a line in this case .
denote by @xmath33 the green current of @xmath0 , see @xcite . with the notation of theorem [ th intro ] , we have that @xmath34 is an invariant probability measure of maximal entropy supported in @xmath2 , see @xcite . under the conditions of theorem [ th intro ]
, we infer that @xmath35 puts no mass on pluripolar sets .
+ a recent work of n. fakhruddin @xcite gives alternate arguments for some of the results in this paper : genericity of non algebraic attracting sets , existence of zariski dense attracting sets of higher codimension .
notice that we do not use the same notion of genericity .
fakhruddin proves that the set of holomorphic endomorphisms of @xmath5 which have no non trivial invariant set ( i.e. not of zero dimension or not @xmath5 ) contains a countable intersection of zariski open sets . on the other hand , we work in a specific family @xmath21 and construct a zariski open set of examples there .
we also observe that the set of endomorphisms in @xmath3 possessing an attracting set of small topological degree is open for the usual topology .
+ the outline of the paper is as follows .
we start with some preliminaries .
then we introduce the notion of small topological degree on an attracting set .
we show that it is a sufficient condition to have a non algebraic attracting set and under this hypothesis , in codimension one , the attracting set is non - pluripolar .
section 3 is devoted to the proof of theorem [ th intro ] . in section 4 , we will exhibit examples in @xmath36 of attracting sets of codimension 2 .
we will see , that the condition of small topological degree is not sufficient in higher codimension to assure that the attracting set is non pluripolar .
using hnon - like maps of small topological degree , we will exhibit the first explicit example of a zariski dense attracting set of codimension 2 .
we finish with some remarks around theorem [ th intro ] and some open questions .
in this section we give some definitions and recall the framework of @xcite .
[ def attracteur ] a set @xmath2 is called an * attracting set * if there exists an open set @xmath37 such that @xmath38 and @xmath39 .
we call such an open set @xmath37 a * trapping region*. if moreover @xmath0 is topologically mixing on @xmath2 then @xmath2 is called an * attractor*. the following proposition is obvious .
let @xmath2 be an attracting set .
then @xmath2 is closed and @xmath40 .
[ def dimension ] an attracting set in @xmath5 is said to be of * dimension * @xmath41 if it supports a positive closed current of bidimension @xmath42 ( i.e. of bidegree @xmath43 ) but none of bidimension @xmath44 with @xmath45 .
* equivalently , an attracting set @xmath46 is said to be of dimension @xmath41 if any ( or one ) trapping region of @xmath2 supports a ( smooth ) positive closed current of bidimension @xmath42 but none of bidimension @xmath44 with @xmath45 . *
if @xmath2 is of dimension @xmath41 then the hausdorff dimension of @xmath2 is at least @xmath47 .
the proof of the following elementary proposition is left to the reader .
[ prop dimension ] if @xmath2 is an algebraic attracting set of dimension @xmath41 then definition [ def dimension ] is equivalent to the classical definition of dimension for algebraic sets .
let @xmath0 be an endomorphism of @xmath5 and let @xmath37 be an open set such that @xmath38 .
we assume that there exists two projective subspaces @xmath48 and @xmath49 of @xmath5 of dimension @xmath50 and @xmath41 , @xmath51 , such that @xmath52 and @xmath53 .
then @xmath54 is of dimension @xmath41 .
in fact , @xmath37 supports the positive closed current @xmath55 $ ] of bidimension @xmath42 but does not support a positive closed current of bidimension @xmath44 with @xmath45 .
indeed , the support of any positive closed current @xmath33 of bidimension @xmath44 with @xmath45 intersects @xmath48 . before stating the main result of @xcite , we need some notation
. let @xmath56 be the projection of center @xmath48 .
more precisely , letting @xmath57 be the projective space , of dimension @xmath58 , containing @xmath48 and passing through a point @xmath59 , then @xmath60 is the unique intersection point between @xmath49 and @xmath57 .
we consider the point @xmath60 as the origin of the complex vector space @xmath61 , where @xmath48 is viewed as the hyperplane at infinity of @xmath62 .
in other words , @xmath63 is viewed as a vector bundle over @xmath49 .
if @xmath64 , we have that @xmath65 . we suppose that @xmath66 if @xmath48 is a point and @xmath49 is a projective hyperplane , i.e. @xmath67 , the previous hypothesis is equivalent to the property that the open subset @xmath68 is star - shaped at @xmath48 .
[ th dinh ] let @xmath0 and @xmath37 be as above .
let @xmath69 be a positive closed @xmath43-form of mass 1 with continuous coefficient and with compact support in @xmath37 .
then the sequence @xmath70 converges to a positive closed current @xmath28 of mass 1 , with support in @xmath54 .
the current @xmath28 does not depend on @xmath69 .
moreover , it is woven , forward invariant ( i.e. under @xmath71 ) and is extremal in the cone of invariant positive closed currents of bidegree @xmath43 with support in @xmath2 . in the sequel
we will refer to @xmath28 as _ the attracting current _ of @xmath2 . here
we focus on the codimension 1 case .
let @xmath0 be an endomorphism of @xmath5 and @xmath72 be the normalised push forward operator , i.e. @xmath73 .
let @xmath33 be a positive closed current of bidegree @xmath29 of mass 1 .
there exists a quasi - psh function @xmath74 , i.e. @xmath74 is locally the difference of a plurisubharmonic function ( psh for short ) and a smooth function , such that @xmath75 .
we call such a function @xmath74 a quasi - potential of @xmath33 .
a dsh function is the difference of two quasi - psh functions , see appendix a.4 . of @xcite for properties of dsh functions and @xcite for basics on psh function .
let @xmath76 be two positive closed currents on @xmath5 of bidegree ( 1,1 ) with the same mass then there exists a dsh function @xmath77 such that @xmath78 .
let @xmath79 be positive closed currents of bidegree ( 1,1 ) with mass 1 .
we denote by @xmath80 the unique dsh function such that @xmath81 and @xmath82 .
[ lemme cv pot ] if @xmath83 converges , in the sense of currents , toward @xmath84 then @xmath85 converges , in @xmath86 and in sense of distributions , toward @xmath80 . by thoerem a.40 of @xcite , @xmath87 depends continuously on @xmath69 . as @xmath88
the proof is complete .
[ lemme forward pot ] let @xmath76 be as above , then @xmath89 and @xmath90 . in the sum @xmath91 , the preimages are counted with multiplicity . as @xmath0 is a holomorphic endomorphism , if @xmath74 is continuous map then @xmath92 is continuous .
we just need to prove this locally in @xmath5 . as @xmath0 is finite ,
if @xmath93 is a small enough open set there exists a psh function @xmath74 , defined on @xmath94 , such that @xmath95 . with the previous notation , @xmath96 is psh on @xmath93 and @xmath97 .
this is classical .
we recall the proof for completeness . up to taking a decreasing regularisation
, we may assume that @xmath74 is smooth .
denote by @xmath98 the critical set of @xmath0 .
as @xmath99 , ( * ? ? ?
* ch.1 theorem ( 2.14 ) ) , and @xmath0 is a submersion on @xmath100 , we have that @xmath101 see ( * ? ? ?
* ch.1 ( 2.15 ) ) .
outside @xmath98 , the map @xmath0 is locally a biholomorphism so @xmath102 is psh on @xmath103 .
moreover , @xmath104 is pluripolar and @xmath92 is continuous on @xmath93 . recall that a locally bounded psh function on @xmath103 admits a unique psh extension on @xmath93 .
thus the equation is true for all @xmath105 .
we now finish the proof of lemma [ lemme forward pot ] .
there exists two psh functions @xmath106 on @xmath37 such that @xmath107 and @xmath108 .
so , on @xmath37 , we have that @xmath109 , thus there exists a pluri - harmonic function @xmath110 such that @xmath111
. then @xmath112 and @xmath113 are psh functions such that @xmath114 and @xmath108 . by the preceding lemma
, we have that @xmath115 on @xmath116 , with @xmath117 .
in this section , we introduce the notion of being of small topological degree on an attracting set in @xmath5 , prove that such an attracting set is never algebraic and that , in codimension 1 , it is non pluripolar .
let @xmath0 be an endomorphism of @xmath5 of algebraic degree @xmath23 and let @xmath37 be a trapping region for an attracting set @xmath2 of dimension @xmath41 .
the endomorphism @xmath0 is said to be of small topological degree on @xmath37 if the number of preimages in @xmath37 of any point belonging to @xmath116 is strictly less than @xmath118 .
the endomorphism @xmath0 is said to be * asymptotically of small topological degree * on @xmath37 if for all @xmath119 we have that @xmath120 .
the notion of being of small topological degree depends on the choice of @xmath37
. we will encounter in section [ section exemples ] examples where @xmath121 is of small topological degree on @xmath116 but not on @xmath37 .
the proof of the following proposition is left to the reader .
the definition of asymptotic small topological degree does not depend on the choice of the trapping region . moreover , if @xmath0 is asymptotically of small topological degree then for each trapping region @xmath37 there exists @xmath122 such that @xmath123 is of small topological degree on @xmath37 .
[ def petit dt ] we say that @xmath0 is of * small topological degree * on an attracting set if @xmath0 is asymptotically of small topological degree on some trapping region . sometimes , we will abbreviate this into attracting set of small topological degree " .
[ cond ouverte ] the property of being of small topological degree on some attracting set is open in the set of endomorphisms of degree @xmath23 .
let @xmath0 be an endomorphism of @xmath5 of small topological degree on some attracting set of dimension @xmath41 and let @xmath37 be a trapping region . replacing @xmath0 by an iterate
we may assume that @xmath0 is of small topological degree on @xmath37 .
then for each @xmath119 , we have that @xmath124 .
let @xmath125 be close to @xmath0 , there exists an open set @xmath93 such that @xmath126 .
thus , the attracting set @xmath127 is of dimension @xmath41 . for each @xmath128
, @xmath129 is close to @xmath130 hence , if @xmath125 is close enough to @xmath0 , we have that @xmath131 .
therefore , @xmath125 is of small topological degree on @xmath93 .
let us prove that such attracting sets are non algebraic .
[ prop non alg ] if a holomorphic endomorphism @xmath0 of @xmath5 is of small topological degree on an attracting set @xmath2 , then @xmath2 is non algebraic .
the proof relies on the following classical lemma .
[ intersection transversale ] let @xmath0 be an endomorphism of @xmath5 and @xmath132 be an invariant algebraic set ( @xmath133 ) of pure dimension @xmath41 , @xmath134 .
then there exists @xmath122 such that @xmath123 fixed all irreducible components of @xmath135 and the topological degree of @xmath136 is @xmath137
. see ( * ? ? ?
* lemma 1.48 ) .
let @xmath41 be the dimension of @xmath2 and let @xmath37 be a trapping region . up to replacing @xmath0 by an iterate , we may assume that @xmath0 is of small topological degree on @xmath37 .
assume that @xmath2 is algebraic .
by proposition [ prop dimension ] , @xmath2 is of dimension @xmath41 .
let @xmath135 be the ( finite ) union of irreducible components of pure dimension @xmath41 of @xmath2
. then @xmath133 , because @xmath40 and @xmath0 does not contract any algebraic subvariety on an algebraic subvariety of lower dimension .
hence , by lemma [ intersection transversale ] , @xmath0 can not be of small topological degree on @xmath135 .
we now proceed with the non pluripolarity ( in codimension 1 ) .
the corresponding result in higher codimension fails ( see below theorem [ th att cont in hyperplan ] ) .
[ th potentiel bourne ] let @xmath0 be a holomorphic endomorphism of @xmath5 and let @xmath2 be an attracting set of dimension @xmath138 .
assume @xmath0 is of small topological degree on some trapping region @xmath37 of @xmath2 .
let @xmath33 be a closed positive current of bidegree ( 1,1 ) of mass 1 , with support in @xmath37 , admitting a bounded quasi - potential .
then each cluster value of @xmath139 has a bounded quasi - potential .
in particular , @xmath2 is non pluripolar . moreover , under the assumptions of theorem
[ th dinh ] , the attracting current @xmath28 has a bounded quasi - potential .
the proof of this theorem is essentially contained in the following lemma .
we recall that @xmath73 .
we denote by @xmath140 the maximum number of preimages in @xmath37 of a point in @xmath116 under @xmath0 .
[ lemme potenitel bourne ] let @xmath79 be positive closed currents of bidegree ( 1,1 ) with mass 1 and with bounded quasi - potentials . for every @xmath141 such that @xmath142 , there exists a constant @xmath143 such that if @xmath144 and @xmath145 then we have that @xmath146 as expected , the proof just consists in making precise the idea that since @xmath80 is pluri - harmonic on @xmath147 , we do not need to focus on preimages belonging to @xmath147 .
assume that we can separate the preimages inside @xmath37 and outside @xmath37 such that the map @xmath113 defined by : @xmath148 is dsh and satisfies @xmath149 .
then we have @xmath150 and we would finish the proof simply by showing that @xmath151 is bounded .
however @xmath152 is not proper , so we actually need to work locally near every point to make this idea work .
let @xmath93 be a small open sets such that @xmath153 may be written as a disjoint union @xmath154 such that @xmath155 , @xmath156 , @xmath157 and the number of preimages ( with multiplicity ) of a point of @xmath93 in @xmath158 is fixed ( and less than @xmath140 ) . in this way , we have @xmath159 and @xmath160 , where @xmath161 as @xmath0 is finite , we may choose for any @xmath119 a neighbourhood @xmath162 small enough such that each connected component of @xmath163 contains a unique preimage of @xmath164 .
( of course @xmath162 can not be uniform with respect to @xmath164 . )
denote by @xmath165 the connected component of @xmath163 which contains @xmath166 .
we denote by @xmath167 the set @xmath168 and by @xmath169 the set@xmath170 . we can reduce @xmath162 such that @xmath171 , @xmath172 and @xmath173 , thus the number of preimages ( with multiplicity ) of a point of @xmath162 which lies in @xmath167 is fixed .
let @xmath174 belonging to @xmath162 .
by lemma [ lemme forward pot ] @xmath175 is a dsh function such that @xmath176 .
we define , @xmath177 by : for all @xmath178 @xmath179 as @xmath180 , @xmath80 is pluri - harmonic outside @xmath141 thus @xmath181 in @xmath162 and @xmath182 . as @xmath183 on @xmath162
, we just have to prove that there exists @xmath184 independent of @xmath79 and @xmath162 such that on @xmath162 @xmath185 we know that @xmath186 is pluri - harmonic on @xmath162 , hence , reducing a little bit @xmath162 if necessary , by the maximum principal there exists @xmath187 such that @xmath188 on @xmath162 .
let us first show that @xmath187 is uniform with @xmath189 and @xmath69 . assume by contradiction that there exists a sequence @xmath83 such that @xmath190 .
we can extract a converging subsequence , still denoted by @xmath83 .
let @xmath191 be its limit . by lemma [ lemme cv pot ]
, the sequence of pluri - harmonic functions @xmath192 converges , in sense of distributions , toward @xmath193 , which is bounded , therefore the convergence is uniform ( see corollary 3.1.4 in @xcite ) .
thus the sequence @xmath194 is bounded , which contradicts the hypothesis .
therefore , there exists @xmath195 independent of @xmath84 such that @xmath196 hence @xmath197 up to shrinking @xmath37 , we may assume that @xmath116 is covered by a finite number of such neighbourhoods @xmath162 .
we infer that there exists @xmath198 such that @xmath199 on @xmath116 .
finally , as @xmath200 is pluri - harmonic outside @xmath201 and @xmath202 , @xmath200 is bounded by a constant @xmath203 on @xmath204 .
thus , we obtain that @xmath205 for similar reasons as in the case of @xmath187 , we may assume that @xmath203 is independent of @xmath79 . therefore , the proof of the lemma is complete . up to replacing @xmath33 by @xmath206 , we assume @xmath207 .
denote by @xmath208 the dsh function such that @xmath209 and @xmath210 , i.e. @xmath211 . then @xmath212 is bounded . by lemma [ lemme potenitel bourne ] , with @xmath213 and @xmath214 , we have @xmath215 it is classical that a sequence @xmath216 satisfying @xmath217 with @xmath218 is bounded . therefore , each cluster value of @xmath219 has a bounded quasi - potential .
if in addition , @xmath0 satisfies then @xmath219 converges toward @xmath28 , thus @xmath28 has also a bounded quasi - potential .
i do not know how to prove the continuity of the quasi - potential .
a reason for this is that the proof does not yield the convergence of the sequence of potentials .
note that the same difficulty appears in dinh s theorem [ th dinh ] .
let @xmath21 be the set of triples @xmath22 of homogeneous polynomials of degree @xmath6 in @xmath13 , such that @xmath14 is the single common zero of @xmath11 . clearly , @xmath21 is a quasi - projective variety . in this section ,
we consider a particular class of endomorphisms of @xmath3 given by the formula @xmath220\rightarrow [ p(z , w):q(z , w):t^d + \varepsilon r(z , w)]\ ] ] where @xmath221 .
attractors for mappings of this form were studied by j.e .
fornss and n. sibony in @xcite , as well as f. rong @xcite .
these endomorphisms preserve the pencil of lines passing through @xmath222 $ ] .
such a line is determined by a point of the form @xmath223 $ ] , thus we identify the line passing through @xmath222 $ ] and @xmath223 $ ] with the point of the line at infinity @xmath223 $ ] .
the dynamics on this pencil of lines is that of @xmath224\rightarrow [ p(z , w):q(z , w)]$ ] on @xmath10 .
[ theoreme ] there exists a zariski open set @xmath225 such that if @xmath25 then for small enough @xmath26 , @xmath27 , the endomorphism @xmath0 of @xmath3 defined by @xmath19\mapsto [ p(z , w):q(z , w):t^d + \varepsilon r(z , w)]\ ] ] admits an attracting set @xmath2 of small topological degree . in particular , @xmath2 is non pluripolar .
moreover , @xmath0 satisfies the condition of theorem [ th dinh ] and the attracting current @xmath28 has a bounded quasi - potential .
since having an attracting set of small topological degree is an open condition , by perturbing in the set of all endomorphisms of @xmath3 , we obtain maps which do not preserve a pencil but also satisfy the conclusion of theorem [ theoreme ] .
the proof gives specific examples in any degree and we can arrange for @xmath0 to be topologically mixing on @xmath2 , therefore we obtain the following : [ prop exist attracteur non pp ] there exists an _ attractor _
@xmath226 of small topological degree .
we now proceed to the proof of theorem [ theoreme ] .
the proof will proceed in four steps .
first we show that any @xmath0 of the form admits an attracting set and satisfies the conditions of theorem [ th dinh ] .
next we give a sufficient condition ensuring that @xmath121 is of small topological degree on some trapping region .
the third step is to show that this condition is algebraic .
finally , we give examples in any degree , therefore showing that the condition is generically satisfied . + * step 1 . *
the map @xmath0 admits an attracting set and satisfies the conditions of theorem [ th dinh ] . +
as @xmath189 is a homogeneous polynomial of degree @xmath23 , there exists @xmath227 such that for all @xmath228 we have that @xmath229 . up to multiplying @xmath26 by @xmath230 we may assume that @xmath231 .
denote by @xmath232 the constant @xmath233 , as @xmath14 is the single common zero of @xmath11 , we have that @xmath234 .
let @xmath235 be the open set of @xmath236\in { { \mathbb{p}}}^2 $ ] such that @xmath237 .
then we have @xmath238 , by definition of @xmath232 . to assume that @xmath239 , all we need is that @xmath240 .
we choose @xmath241 .
we let the reader check that the choice @xmath242 is convenient .
thus @xmath243 is an attracting set .
denote by @xmath244}$ ] the line passing through @xmath222 $ ] and @xmath223 $ ] and by @xmath245 the line at infinity .
the line @xmath244}$ ] intersects @xmath246 into an open disc centered at @xmath223 $ ] .
thus , by taking @xmath247 \rbrace$ ] , @xmath248 and @xmath249 , @xmath0 satisfies the hypothesis . +
* step 2 . * a sufficient condition for being of small topological degree .
+ here we will introduce for @xmath250 subsets @xmath251 and @xmath252 of @xmath253 and a condition on these sets insuring that for all @xmath254 small enough @xmath121 is of small topological degree on @xmath255 , where @xmath0 is defined by , ( see proposition [ prop petit deg top ] ) . throughout the proof , by a line we mean a line passing through @xmath222 $ ] , unless for the line at infinity @xmath245 . recall that the image and the preimages of a line is a line and that @xmath256})=l_{f_\infty({[z : w]})}$ ] .
for the sake of convenience , we will confuse the line @xmath244}$ ] with the point @xmath17 $ ] in @xmath245 and denote by @xmath257 the action of the lines .
so @xmath251 and @xmath252 can be seen as a set of points in @xmath245 or a set of lines .
the attracting set we obtain is a complex version of the solenoid .
the difference is that the `` branches '' of @xmath255 must necessarily cross , so the map can not be injective .
the proof consists in analysing the geometry of @xmath255 , as well as its self - crossings , and the behaviour of the preimages , such that estimating the number of preimages staying in @xmath255 reduces to a combinatoric problem . for maps of the form
, a generic line has @xmath23 preimages and a generic point has @xmath258 preimages ( in @xmath3 ) , with @xmath23 in each line .
so there are two ways to control the number of preimages ( which lie in @xmath255 ) of a point : * controlling the number of lines that contain preimages staying in @xmath246 .
* for each points in such a line , bounding the number of preimages that belong to @xmath255 .
we leave the reader check that the image of a disc in @xmath244}$ ] centered at @xmath223 $ ] is a disc in @xmath259)}$ ] centered at @xmath260)$ ] . a line @xmath261 having @xmath23 preimages ( with multiplicity ) , @xmath262 is made of at most @xmath23 discs which might intersect .
denote by @xmath263 the set of lines @xmath261 such that @xmath262 is made of @xmath23 disjoint discs .
thus if @xmath264 , the preimages of a point @xmath265 , which lie in @xmath246 , are contained in a single line @xmath266 , i.e. @xmath267 . through the end of the proof , for each @xmath268 , we normalize @xmath269 so that @xmath270 the line @xmath244}$ ] is in @xmath263 iff for @xmath271 in @xmath272 and @xmath273 such that @xmath274,[z_j : w_j : t_j]\in u_\rho$ ] we have that @xmath275 . by definition of @xmath246 , for this , it suffices that @xmath276 $ ] all have the same image and are either all in @xmath246 , or all outside @xmath246 .
therefore , for any line @xmath261 and any point @xmath277 , if we fix a preimage @xmath244}$ ] of @xmath261 , the point @xmath164 has @xmath23 preimages ( with multiplicity ) in the line @xmath244}$ ] , they are of the form @xmath278 $ ] .
thus , @xmath164 has 0 or @xmath23 preimages ( with multiplicity ) in @xmath244}\cap u_\rho$ ] . to further reduce the number of preimages
we must see if they are or not in @xmath255 .
denote by @xmath279 the set of lines @xmath244}$ ] such that for all @xmath237 there exists at most one @xmath280 such that @xmath278 \in f(u_\rho)$ ] and denote by @xmath281 the set of lines whose preimages all of lie in @xmath279 .
the line @xmath244}$ ] is in @xmath279 iff for all @xmath282 , all @xmath273 such that @xmath274,[z_j : w_j : t_j]\in u_\rho$ ] and all @xmath283 we have that @xmath284 . for this
, it suffices that for each @xmath283 @xmath285 and @xmath286 , for any preimage @xmath266 of @xmath261 , @xmath164 has at most one preimage in @xmath287 . if in addition @xmath288 then a point @xmath286 has a single preimage in @xmath255 .
+ we need to understand the complement of @xmath289 .
+ denote by @xmath290}, .. ,l_{[z_d : w_d]}$ ] the preimages ( with possible repetitions ) of a line @xmath244}$ ] .
a sufficient condition for a line @xmath244}$ ] not to be in @xmath263 is that there exists @xmath271 in @xmath272 such that @xmath291 ) = f([z_j : w_j:0])$ ] .
denote by @xmath292 the set of lines @xmath293 } ,
l_{[z_j : w_j]}$ ] which satisfy this condition and let @xmath294 .
therefore , @xmath244 } \in x$ ] iff there exists @xmath271 in @xmath295 such that @xmath296}\cap u_\rho ) = f(l_{[z_j : w_j]}\cap u_\rho)$ ] .
a point in @xmath255 is of the form @xmath297 $ ] with @xmath237 .
a sufficient condition for a line @xmath244}$ ] not to be in @xmath279 is that there exists @xmath282 and @xmath283 such that @xmath298 = [ p(z_j , w_j):q(z_j , w_j):\varepsilon { { \textit{\emph{e}}}}^{\frac{2ik\pi}{d } } r(z_j , w_j)]$ ] .
possibly @xmath299 in which case this correspond to @xmath300 and necessarily , as in @xmath244}\cap u_\rho$ ] , there will be points in @xmath256})$ ] that have @xmath23 preimages in @xmath244}\cap f(u_\rho)$ ] .
thus , two problems can occur ( when @xmath299 or not ) .
denote by @xmath301 the set of lines @xmath293 } , l_{[z_j : w_j]}$ ] which satisfy this condition and let @xmath302 and @xmath303 . by definition of @xmath251 and @xmath252 , @xmath304}\in x & \longleftrightarrow & \exists i\neq j
\in \lbrace 1, .. ,d \rbrace \text { such that } \\ & & r(z_i , w_i ) - r(z_j , w_j)=0 \longleftrightarrow l_{[z_i : w_i ] } , l_{[z_j : w_j ] } \in x_{-1 } \nonumber\\ \label{def y }
f(l_{[z : w]})\in y & \longleftrightarrow & \exists i , j\in \lbrace 1, .. ,d \rbrace , \ , k\in \lbrace 1, .. ,d-1 \rbrace \text { such that } \\ & & r(z_i , w_j ) - { { \textit{\emph{e}}}}^{\frac{2ik\pi}{d } } r(z_j , w_j)=0 \longleftrightarrow l_{[z_i : w_i ] } , l_{[z_j : w_j ] } \in y_{-2}.\nonumber\end{aligned}\ ] ] by definition , we have that @xmath305 .
in fact , @xmath306 ( resp .
@xmath307 ) precisely when @xmath261 is not in @xmath263 ( resp .
@xmath281 ) for any @xmath26 .
denote by @xmath308 the map @xmath309)=[p(z ,
w ) : q(z , w):r(z , w)]$ ] and by @xmath310 the map @xmath311 . as @xmath312 ( resp .
@xmath301 ) iff @xmath313 ( resp .
@xmath301 ) , we will , sometimes , identify @xmath314 with @xmath315 , particularly in the examples .
we note for further reference that , by identifying @xmath244}$ ] to @xmath17\in { { \mathbb{p}}}^1 $ ] , @xmath292 and @xmath301 may be written formally as : @xmath316 the next proposition is a sufficient condition to be of small topological degree on the attracting set @xmath2 .
recall that @xmath317 are sets of lines passing through @xmath222 $ ] , that we identify such a line with a point in @xmath253 and that the dynamics of the pencil of these lines is the same as that of @xmath318\rightarrow [ p(z , w):q(z , w)]$ ] .
note that @xmath319 only depend on the choice of @xmath30 , not on @xmath26 .
[ prop petit deg top ] denote by @xmath320 and @xmath321 the sets @xmath322 and @xmath323 . if @xmath30 are such that 1 .
[ cond ens .
distincts]@xmath324 and @xmath325 are disjoint 2 .
[ cond pt qui retombe ] @xmath326 , or equivalently @xmath327 then for small enough @xmath32 there exists @xmath328 such that @xmath121 is of small topological degree on @xmath255 , with @xmath0 define by . by , it is clear that @xmath329 are non empty algebraic set .
moreover , if the condition [ cond ens .
distincts ] of proposition [ prop petit deg top ] is satisfied , then @xmath329 are finite .
see section [ section version simple ] for a simplified version of these conditions . for @xmath26 as before , we fix @xmath330 , as in , so that @xmath239 .
denote by @xmath331 ( resp .
@xmath332 ) the @xmath333-neighbourhood of @xmath251 ( resp .
@xmath334 ) in @xmath245 .
we choose @xmath333 small enough such that @xmath331 ( resp .
@xmath332 ) is the union of disjoint open discs centered at the points of @xmath251 ( resp .
@xmath334 ) and such that we have that : * @xmath335 , by hypothesis , * @xmath336 , since @xmath337 , * and @xmath338 , by hypothesis . fix @xmath328 and @xmath333 as above .
then if @xmath254 is small enough , we have that @xmath339 indeed , the first inclusion is due to the definition of @xmath251 and @xmath252 . regarding the second inclusion , recall that @xmath290 } , .. , l_{[z_d : w_d]}$ ] denoted the preimages ( with possible repetitions ) of a line @xmath244}$ ] . by and , since @xmath340 is compact and @xmath341 and @xmath342 , there exists @xmath343 such that if @xmath293 } , l_{[z_j : w_j ] } \notin f^{-1}(x^r)$ ] with @xmath271 then @xmath344 and if @xmath293 } , l_{[z_j : w_j ] } \notin f^{-2}(y^r)$ ] , ( @xmath345 can be equal here ) then for each @xmath283 @xmath346 since @xmath330 , @xmath6 and @xmath347 is independent of @xmath23 and @xmath26 , if @xmath26 is small enough then @xmath348 , which implies that if @xmath244}\notin x^r$ ] and @xmath345 are distinct then latexmath:[\[\label{ineg v0 } and if @xmath244}\notin f_\infty^{-1}(y^r)$ ] , ( @xmath345 can be equal here ) then for all @xmath283 @xmath350 if @xmath244}\notin x^r$ ] then for all @xmath345 , @xmath293 } , l_{[z_j : w_j ] } \notin f_\infty^{-1}(x^r)$ ] with @xmath271 , thus , from and , @xmath244}\in a$ ] .
if @xmath244}\notin f_\infty^{-1}(y^r)$ ] then for all @xmath345 , we have that @xmath293 } , l_{[z_j : w_j ] } \notin f_\infty^{-2}(y^r)$ ] , thus , from and , @xmath244}\in b_{-1}$ ] .
consequently , if @xmath244}\notin y^r$ ] , no preimage of @xmath244}$ ] is in @xmath351 , thus all preimages of @xmath244}$ ] are in @xmath279 , i.e. @xmath244}\in b$ ] .
therefore , if @xmath244}\notin z^r$ ] then @xmath244}\in a\cap b$ ] , i.e. @xmath352 .
this finish the proof of the lemma .
to finish the proof of proposition [ prop petit deg top ] we will prove that a point @xmath353 has at most @xmath258 preimages under @xmath121 ( which lie in @xmath354 ) .
denote by @xmath355\rbrace\rightarrow l_\infty $ ] the projection defined by @xmath356)=[z : w]$ ] . the proof is summarized the following diagram : + + a point in @xmath357 has one preimage in @xmath358 .
moreover , the complementary of @xmath281 is contained in @xmath359 and the complement of @xmath263 is contained in @xmath331 . since @xmath335
, we infer that a point in @xmath360 has at most @xmath23 preimages in @xmath358 .
thus , a point has at most @xmath23 preimages under @xmath0 in @xmath358 .
+ for the sake of simplicity , until the end of the proof , without precision by preimage we mean preimage in @xmath358 .
pick a point @xmath353 . + case 1 . if @xmath361 then @xmath362 , so @xmath164 has one preimage under @xmath0 and @xmath363 .
+ case 2 . if @xmath364 then @xmath164 has at most @xmath23 preimages under @xmath0 . for each preimage of @xmath164
we have two cases : + case 2.1 .
either it is still in @xmath365 .
thus it has @xmath23 preimages under @xmath0 which are in @xmath366 ( because @xmath336 ) , but since @xmath367 , these preimages leave @xmath365 and thus have only one preimage under @xmath0 .
it follows that this preimage of @xmath164 has at most @xmath23 preimages under @xmath368 .
+ case 2.2 . or it is outside @xmath365 , then it has one preimage .
it follows that this preimage of @xmath164 has at most @xmath23 preimages under @xmath368
. therefore , @xmath369 in all cases . consequently
, @xmath121 is of small topological degree on @xmath255 .
this concludes the proof of the proposition .
denote by @xmath370 the subset of @xmath30 satisfying the hypotheses of proposition [ prop petit deg top ] .
+ at this point to finish the proof of theorem [ theoreme ] it only remains to prove that @xmath370 is a zariski open set and that it is not empty . +
* step 3 . *
the subset @xmath370 of @xmath371 is a zariski open set .
+ this is simply an exercise in complex geometry .
recall from that @xmath372 , @xmath373 , @xmath322 and @xmath323 , where @xmath374 . as already observed ,
@xmath375 only depend on the choice of @xmath30 and not on the choice of @xmath254 . recall that the conditions of proposition [ prop petit deg top ] are 1 .
@xmath251 and @xmath325 are disjoint 2 .
@xmath326 or equivalently @xmath376 .
set @xmath377 and denote by @xmath378 the map @xmath379 and by @xmath380 the sets @xmath381 , @xmath382 and @xmath383 . thus @xmath384 is an algebraic subvariety of @xmath385 as well as @xmath386 .
therefore , @xmath387 and @xmath388 are algebraic subvarieties in @xmath389 , and @xmath390 and @xmath391 are algebraic subvarieties too .
the complementary of @xmath370 is in the image of @xmath392 by the projection @xmath393 , thus it is a zariski closed subset in @xmath21 . + * step 4 . *
the zariski open set @xmath370 defined in step 3 is not empty . + let us show that @xmath394 $ ] satisfies the hypotheses of proposition [ prop petit deg top ] for almost every @xmath395 small enough .
we first assume @xmath396 .
if @xmath244}$ ] is a preimage of a line , the other preimages are @xmath397}$ ] with @xmath283 . for @xmath398 put @xmath399.\ ] ] we let the reader check that @xmath400,[1:0 ] \rbrace$ ] and @xmath401 ; l\in \lbrace 1, .. ,d-1 \rbrace , k\in \lbrace 0, .. ,d-1 \rbrace \rbrace\\ \nonumber & = \lbrace [ 0:1]\rbrace \bigcup \lbrace z_{k , l } ; l\in \lbrace 1, .. ,d-1 \rbrace , k\in \lbrace 0, .. ,d-1 \rbrace , k+l\neq d \rbrace .\end{aligned}\ ] ] thus @xmath402,[a:1 ] \rbrace , \\ & y & = f_\infty^2(y_{-2 } ) = \lbrace f_\infty([1:b])\rbrace \bigcup \lbrace f_\infty^2(z_{k , l } ) ; l\in \lbrace 1, .. ,d-1 \rbrace , k\in \lbrace 0, .. ,d-1 \rbrace , k+l\neq d \rbrace .
\end{aligned}\ ] ] suppose that @xmath403 are small and @xmath404 . if @xmath405 then @xmath406 $ ] , @xmath407 $ ] and @xmath408\approx [ a:1]$ ] , otherwise @xmath409 and @xmath410 . if @xmath411 then @xmath412 @xmath413 and @xmath414 are far from @xmath415,[1:0]$ ] . in particular , @xmath416,[1:0],[1:b],[a:1],[b^d+a : b^{d+1}+1],[1+a^{d+1}:b+a^d ] \rbrace = \emptyset$ ] . we conclude that , if @xmath403 are small and @xmath404 then @xmath251 and @xmath325 are disjoint , i.e. @xmath0 satisfies the hypothesis of proposition [ prop petit deg top ] . to check assumption ,
\bigcup \lbrace f_\infty^2(z_{k , l } ) ; k+l\neq d \rbrace,\ ] ] so that @xmath418,[1:0],[1:{{\textit{\emph{e}}}}^{\frac{2in\pi}{d } } b ] , { { \textit{\emph{e}}}}^{\frac{2in\pi}{d}}f_\infty(z_{k , l } ) | n\in { { \mathbb n}}\rbrace.\ ] ] the @xmath412 are independent of @xmath419 , so we can choose @xmath419 so that @xmath420 , thus @xmath421 \rbrace$ ] . since @xmath422)=\lbrace [ 0:1]\rbrace \notin z$ ] , we conclude that @xmath0 satisfies the second assumption of proposition [ prop petit deg top ] .
this completes the proof of theorem [ theoreme ] .
we now prove proposition [ prop exist attracteur non pp ] .
we first show that @xmath423 $ ] satisfies the hypotheses of proposition [ prop petit deg top ] .
the preimages of @xmath424}$ ] are @xmath425}$ ] and @xmath426}$ ] .
we let the reader check that @xmath427,[1:0 ] \rbrace$ ] and @xmath428),f_\infty ( [ 1:9])\rbrace=\lbrace [ 9:1],[17 ^ 2:1 ] \rbrace$ ] , which are disjoint .
thus @xmath429,[1:0],[9:1],[17 ^ 2:1 ] \rbrace$ ] and @xmath430,[0:1],[1:2],[1:-1],[1:9],[1:8 ] \rbrace$ ] , therefore @xmath431,[1:9 ] \rbrace$ ]
. finally , @xmath432,[1:2],[1:-1 ] \rbrace $ ] and @xmath0 satisfies the hypotheses of proposition [ prop petit deg top ] .
thus , by proposition [ th potentiel bourne ] and [ prop petit deg top ] , @xmath0 has an attracting set @xmath226 supporting a positive closed current of bidegree ( 1,1 ) which admits a bounded quasi potential . from the work of @xcite lemma 2.12 and @xcite corollary 4.15 ,
the natural extension @xmath433 of @xmath257 is semi - conjugate to @xmath0 restricted to @xmath2 . since @xmath434
, @xmath257 is topologically mixing on @xmath245 , thus @xmath0 is topologically mixing on @xmath2 and @xmath2 is an attractor .
in this part , we will explain why the condition of being of small topological degree is not enough to ensure non pluripolarity in higher codimension .
for this we exhibit attracting sets of small topological degree in @xmath36 ( of codimension 2 ) which are contained in a hyperplane ( see theorem [ th att cont in hyperplan ] ) .
we also construct the first explicit example of a zariski dense attracting set of higher codimension ( see theorem [ th ex z dense ] ) . [ th att cont in hyperplan ] for a generic choice of @xmath30 , there exists @xmath435 such that for all @xmath436 and @xmath437 , the mapping @xmath438 & \mapsto & [ p(z , w ) : q(z , w):t^d+\varepsilon_1 r(z , w):u^d+\varepsilon_2 q(z , w ) -\varepsilon_2^d w^d ] \\
\end{array}\ ] ] is of small topological degree on an attracting set @xmath2 of dimension @xmath439 which is contained in @xmath440;u=\varepsilon_2 w \rbrace\simeq { { \mathbb{p}}}^2 $ ] . in particular
, @xmath2 is non algebraic nevertheless it is contained in a hyperplane . if @xmath441 \mapsto [ p_0(z , w , t ) : p_1(z , w , t):p_2(z , w , t)]$ ] admits an attracting set @xmath2 of dimension 1 then @xmath442 \ , |\ , [ z :
w : t]\in { \mathscr{a}},u=0 \rbrace$ ] is an attracting set for @xmath443 \mapsto [ p_0(z , w , t ) : p_1(z , w , t):p_2(z , w , t):u^d]$ ]
. but @xmath444 is not of small topological degree on @xmath445 even if @xmath0 is of small topological degree on @xmath2 .
this is why we need a non - trivial construction in theorem [ th att cont in hyperplan ] .
denote by @xmath446 the maps @xmath447\mapsto[p(z , w ) : q(z , w):t^d+\varepsilon_1 r(z , w)],\\ f_2:&[z : w : u]\mapsto[p(z , w ) : q(z , w ) : u^d+\varepsilon_2 q(z , w ) -\varepsilon_2^d w^d ] \end{array}\ ] ] and denote by @xmath448 the open sets @xmath449;|t|<\rho_1\max(|z|,|w| ) \rbrace , u_2=\lbrace [ z : w : u];|u-\varepsilon_2 w|<\rho_2\max(|z|,|w| ) \rbrace .\ ] ] first we choose @xmath22 in the non empty zariski open set of @xmath21 such that the map @xmath450 satisfies the hypotheses of proposition [ prop petit deg top ] and fix @xmath451 small enough and @xmath452 such that @xmath453 , see step 1 of the proof of theorem [ theoreme ] . by modifying the end of step 2 of the proof of theorem [ theoreme ] , a point in @xmath454
has at most @xmath455 preimages in @xmath456 under @xmath457 .
it is clear that the hyperplane @xmath440;u=\varepsilon_2 w \rbrace$ ] is invariant under @xmath0 and we see that the dynamics on it is the same as that of @xmath450 .
denote by @xmath458 the constant @xmath459 , with @xmath143 , and by @xmath37 the set @xmath460 ; |t|<\rho_1 \max(|z|,|w| ) , |u-\varepsilon_2 w|<\rho_2 \max(|z|,|w| ) \right\rbrace.\ ] ] we choose @xmath461 small enough such that @xmath462 and we will fix @xmath184 later .
we let the reader check that @xmath38 and denote by @xmath54 , hence @xmath2 is an attracting set of dimension 1 .
we now prove that , for a generic choice of @xmath30 , @xmath463 is of small topological degree on @xmath116 . let us further assume that for all @xmath464 @xmath465)\neq [ 1:0]$ ] or equivalently @xmath466\notin f_\infty^{-i}([1:0])$ ] where @xmath467 is defined by @xmath468)=[p(z , w):q(z , w)]$ ] . arguing as in step 3 of the proof of theorem [ theoreme ] , we see that this condition is algebraic . to show that it is generically satisfied , we need to find an example .
let us give an example .
we know by step 4 of the proof of theorem [ theoreme ] that for almost every @xmath469 with @xmath396 small enough @xmath470\mapsto [ w^d+a z^d : b w^d+z^d : t^d + \varepsilon_1 ( z^d+z^{d-1}w)]$ ] satisfies the hypotheses of proposition [ prop petit deg top ] . for all @xmath471 such that @xmath472 , we have that for each @xmath473 @xmath474)\neq [ 1:0]$ ] , because if @xmath475 are non negative real numbers then @xmath476 are positive .
therefore , it is a generic condition .
fix @xmath477 a disc centered at @xmath466 $ ] in @xmath10 such that for all @xmath464 @xmath478 , fix @xmath143 such that if @xmath17\notin { \text{d}}$ ] then for each @xmath479 @xmath480 this implies that if @xmath481\in u_{2}$ ] and @xmath17\notin { \text{d}}$ ] then for each @xmath479 , @xmath482\notin u_{2}. \nonumber\ ] ] let @xmath481\in f^2_2(u_{2})$ ] and @xmath483 $ ] such that @xmath484 $ ] . if @xmath483\notin { \text{d}}$ ] there exists a unique @xmath485 , such that @xmath486)=[z : w : u]$ ] and @xmath487\in f_2(u_{2})$ ]
otherwise , if @xmath483\in { \text{d}}$ ] there exists at most @xmath23 such values @xmath485 .
thus , if we fix @xmath488 $ ] such that @xmath489)=[z : w]$ ] , since for all @xmath464 @xmath478 , there exists at most one @xmath490 such that @xmath491)\in { \text{d}}$ ] .
so there exists at most @xmath23 values @xmath492 , such that @xmath493)=[z : w : u]$ ] and @xmath494\in f_2 ( u_{2})$ ] .
let @xmath495\in f^7(u)$ ] , then @xmath31\in f_{1}^7(u_{1})$ ] , so there exists at most @xmath455 preimages @xmath496 $ ] in @xmath456 such that @xmath497)=[z : w : t]$ ] . for each choice
@xmath488 $ ] , there exists at most @xmath23 values @xmath492 , such that @xmath493)=[z : w : u]$ ] and @xmath494\in f_2 ( u_{2})$ ] .
then for each choice @xmath496\in f_1(u_1)$ ] , there exists at most @xmath23 values @xmath492 , such that @xmath498)=[z : w : t : u]$ ] and @xmath499\in f ( u)$ ] . in conclusion , a point in @xmath500 has at most @xmath501 preimages under @xmath463 in @xmath116 .
thus @xmath463 is of small topological degree on @xmath116 .
we now turn to the zariski dense example of codimension 2 in @xmath36 .
we first recall some results about hnon - like maps .
let @xmath502 be an open set and @xmath477 an open disc , denote by @xmath503 and @xmath504 the sets @xmath505 and @xmath506 , @xmath507 , @xmath508 a neighbourhood of @xmath503 and @xmath140 the maximum number of preimages in @xmath503 of a point in @xmath509 .
the map @xmath510 is called a horizontal - like mapping if 1 .
@xmath511 2 .
@xmath512 , if , furthermore , @xmath0 is injective , @xmath0 is said to be hnon - like .
see @xcite for some proprieties of hnon - like and horizontal - like mappings .
a current @xmath33 is said to be horizontal if its support is contained in @xmath513 .
let @xmath0 be a horizontal - like map .
then there exists an integer @xmath514 such that for every normalized horizontal positive closed current @xmath33 in @xmath503 , @xmath515 is normalized , horizontal , positive and closed in @xmath503 .
the integer @xmath23 is called the degree of @xmath0 see @xcite or @xcite .
[ th pot holder ] let @xmath0 be a hnon - like map in @xmath503 .
if @xmath33 is an invariant current by @xmath516 , then the potentials of @xmath33 are hlder continuous .
this is identical to theorem 2.12 in @xcite .
let @xmath0 be a horizontal - like map of degree @xmath23 .
if the number of preimages in @xmath503 of a point in @xmath517 is strictly less than @xmath23 , @xmath0 is said to be of small topological degree .
[ lemme henon ] for @xmath254 small enough , there exists @xmath328 such that the map @xmath518\mapsto[z^2 + 0.1 w^2:w^2:t^2+\varepsilon(z^2+zw)]\ ] ] is a hnon - like map on @xmath519 where @xmath520;0.8|w|<|z|<1.2|w|,|t|<\rho \max(|z|,|w| ) \rbrace $ ] .
furthermore , denoting by @xmath521 , a point in @xmath522 has a unique preimage in @xmath2 and this preimage is still in @xmath522 .
denote by @xmath246 the set@xmath523;|t|<\rho \max(|z|,|w| ) \rbrace$ ] , by the above , it follows that for @xmath26 small enough , there exists @xmath328 such that @xmath524 .
denote by @xmath502 the set @xmath525;0.8|w|<|z|<1.2|w|\rbrace$ ] , an easy computation shows that @xmath526 , where @xmath527\mapsto[z^2 + 0.1 w^2:w^2]$ ] . thus @xmath0 is a horizontal - like map on @xmath503 . with notation as in step 2 of theorem [ theoreme ]
, we get that @xmath528,[1:0 ] \rbrace$ ] and @xmath529 $ ] . reducing @xmath26 if necessary , we get that @xmath530 ( see step 2 of theorem [ theoreme ] ) , thus @xmath0 is injective on @xmath519 . as @xmath0 is hnon - like on @xmath519 , if @xmath236\in { \mathscr{a}}\cap { \mathbb{b}}$ ] then @xmath236\in { \mathscr{a}}\cap f({\mathbb{b}})$ ] and @xmath17\in w$ ] thus there exists a unique @xmath531\in { \mathscr{a}}$ ] such that @xmath532)=[z : w : t]$ ] .
moreover , as @xmath533 , @xmath483 $ ] is still in @xmath502 .
[ th ex z dense ] there exists @xmath534 such that the map @xmath535\mapsto[z^2 + 0.1 w^2:w^2:t_1 ^ 2+\varepsilon_1(z^2+zw):t_2 ^ 2+\varepsilon_2 ( z^2+zw)]\ ] ] admits a zariski dense attracting set @xmath2 close to the line @xmath536 ; t_1=0 ; t_2=0 \rbrace$ ] .
denote by @xmath537 the projections defined by @xmath538)=[z : w : t_1]$ ] and @xmath539)=[z : w : t_2]$ ] .
for @xmath540 denote by @xmath541 the map @xmath542\mapsto[z^2 + 0.1 w^2:w^2:t_i^2+\varepsilon_1(z^2+zw)]$ ] , by @xmath543 and @xmath544 the sets @xmath545;|t_i|<\rho_i\max(|z|,|w| ) \rbrace$ ] and @xmath546;0.8|w|<|z|<1.2|w|,|t_i|<\rho_i \max(|z|,|w| ) \rbrace $ ] .
we choose @xmath547 and @xmath548 such that @xmath549 .
denote by @xmath550 . and by @xmath551 the positive closed current given by theorem [ th dinh ] for @xmath541 . by theorem [ th pot holder ] and lemma [ lemme henon ] , there exists @xmath552 such that , in any neighbourhood @xmath553 of a point in @xmath554 , @xmath555 has hausdorff dimension at least @xmath556 . on the other hand , j.e .
fornss and n. sibony ( * ? ? ?
2.17 ) proved that if @xmath557 is small enough then @xmath558 has hausdorff dimension less than @xmath559 .
denote by @xmath37 the set @xmath560;|t_1|<\rho_1\max(|z|,|w|),|t_2|<\rho_2\max(|z|,|w| ) \rbrace$ ] , thus @xmath38 .
the attracting set @xmath54 is of dimension 1 , i.e. of codimension 2 .
we have that @xmath561 and @xmath562 , thus @xmath2 is non algebraic .
we will use the discrepancy between the hausdorff dimension of @xmath563 and @xmath558 to show that @xmath2 must be zariski dense .
denote by @xmath564 the natural extension of @xmath565 where @xmath318\mapsto [ p(z , w):q(z , w)]$ ] , see section 3 of @xcite for an introduction .
let @xmath566\in { { \mathbb{p}}}^1 $ ] and denote by @xmath567 the line ( resp .
hyperplane ) passing through @xmath222 $ ] and @xmath223 $ ] ( resp .
@xmath568 $ ] , @xmath569 $ ] and @xmath570 $ ] ) .
we have already seen that @xmath571 is a disk and its image is relatively compact in @xmath572 , see step 1 of theorem [ theoreme ] .
thus if @xmath573 is a sequence of preimages of @xmath164 then @xmath574 is a unique point in @xmath575 .
therefore , this defines a continuous and onto map @xmath576 , see ( * ? ? ?
4.13 ) or ( * ? ? ?
* lemma 2.8 ) .
we have that if @xmath577 $ ] then @xmath578)$ ] .
we can also define a continuous and onto map @xmath579 by replacing disks by balls in the previous argument .
this map satisfies @xmath580 , for @xmath581 .
we have the three commutative diagrams : @xmath582^{\hat{f}_\infty } \ar[d]^{\pi_1 } & \hat{{{\mathbb{p}}}}^1\ar[d]^{\pi_1}\\ { \mathscr{a}}_1 \ar[r]^{f_1 } & { \mathscr{a}}_1 \\ } \qquad \qquad \xymatrix { \hat{{{\mathbb{p}}}}^1 \ar[r]^{\hat{f}_\infty } \ar[d]^{\pi_2 } & \hat{{{\mathbb{p}}}}^1\ar[d]^{\pi_2}\\ { \mathscr{a}}_2 \ar[r]^{f_2 } & { \mathscr{a}}_2 \\ } \qquad \qquad \xymatrix { \hat{{{\mathbb{p}}}}^1 \ar[r]^{\hat{f}_\infty } \ar[d]^\pi & \hat{{{\mathbb{p}}}}^1\ar[d]^\pi\\ { \mathscr{a}}\ar[r]^{f } & { \mathscr{a}}\\ } \ ] ] by lemma [ lemme henon ] , @xmath541 is hnon - like on @xmath583 thus @xmath584 restricted to @xmath585 and @xmath586 are homeomorphisms , where @xmath587;0.8|w|<|z|<1.2|w|,|t_1|<\rho_2 \max(|z|,|w| ) , |t_2|<\rho_2 \max(|z|,|w| ) \rbrace $ ] .
moreover , as @xmath578)$ ] and @xmath588)$ ] , we have that @xmath589 .
suppose that the zariski closure of @xmath2 is a hypersurface @xmath135 .
as @xmath590 is non algebraic , @xmath135 can not be the union of hyperplane passing through @xmath569 $ ] let @xmath591 be an open set contained in @xmath592 such that @xmath593 is the union of biholomorphic copies of @xmath591 included in the regular points of @xmath135 .
at least one of these copies intersect @xmath2 in a set of hausdorff dimension at least @xmath594 , because @xmath595 has hausdorff dimension at least @xmath594 .
denote it by @xmath596 and denote by @xmath597 the irreducible component of @xmath135 which contains @xmath596 .
if @xmath598 is not a curve then , up to reducing @xmath591 , @xmath599 is a biholomorphism .
thus @xmath600 , but this contradicts @xmath601 and @xmath602 . otherwise , let @xmath603 and @xmath502 be an open set such that @xmath604 . as @xmath605 is continuous there
exists an open set @xmath606 such that @xmath607 and @xmath608 . moreover , @xmath609 so @xmath610 .
then @xmath611 is included in the curve @xmath598 but this contradicts the fact that @xmath612 .
we have reached a contradiction .
therefore , @xmath2 is zariski dense in @xmath36 . for each @xmath613
the point @xmath466 $ ] has multiplicity @xmath614 for @xmath615 thus @xmath0 is not of small topological degree on @xmath2 .
the original generic condition that we had in mind for small topological degree map on an attracting set ( proposition [ prop petit deg top ] ) was somewhat simpler .
unfortunately , we were only able to find corresponding examples in degree 2 .
[ cond pt qui retombe]. @xmath616 .
+ it is clear from the proof of proposition [ prop petit deg top ] that this implies that for all @xmath617 small enough there exists @xmath328 such that @xmath368 is of small topological degree on @xmath255 . the map @xmath618 $ ] satisfies the hypothesis [ cond ens .
distincts ] of the proposition [ prop petit deg top ] and [ cond pt qui retombe] for almost every @xmath469 .
in fact , we have that @xmath400,[1:0 ] \rbrace$ ] thus @xmath619,[a:1 ] \rbrace$ ] and @xmath620,[i:-1],[1:-1 ] \rbrace$ ] thus @xmath621,[(b-1)^2+a(1-a)^2:b(b-1)^2+(1-a)^2 ] \rbrace.\ ] ] therefore , for almost every @xmath469 , @xmath324 and @xmath325 are disjoint , i.e. @xmath0 satisfies the hypothesis [ cond ens . distincts ] of [ prop petit deg top ] and as @xmath622 @xmath623,[a:1],[(1+b)^2+a(1+a)^2:b(b+1)^2+(1+a)^2],[(b-1)^2+a(1-a)^2:b(b-1)^2+(1-a)^2 ] \rbrace\ ] ] and @xmath624,[1:0],[1+a : b+1],[1-a : b-1],[1+a :- b-1],[1-a :- b+1 ] \rbrace,\ ] ] for almost every @xmath469 , we have that @xmath625 , thus @xmath0 satisfies the condition [ cond pt qui retombe]. to extend theorem [ theoreme ] to higher dimension , we consider the family of endomorphisms of @xmath5 of the form @xmath626\mapsto [ p_0(z_0: .. :z_{k-1}): .. :p_{k-1}(z_0: .. :z_{k-1}):z_k^d+\varepsilon p_{k}(z_0: .. :z_{k-1})],\ ] ] where @xmath627 are homogeneous polynomials of degree @xmath6 such that @xmath628 is the single common zero of @xmath629 . let @xmath630 ; |z_k|<\rho \max ( |z_0|, .. ,|z_k|)\rbrace$ ] , then if @xmath631 are small enough we have that @xmath524 .
denote by @xmath257 the map @xmath632\mapsto [ p_0(z_0: .. :z_{k-1}): .. :p_{k-1}(z_0: .. :z_{k-1})],\ ] ] @xmath633\mapsto [ p_0(z_0: .. :z_{k-1}): .. :p_{k-1}(z_0: .. :z_{k-1}):\varepsilon p_{k}(z_0: .. :z_{k-1})].\ ] ] we define @xmath252 as before proposition [ prop petit deg top ] and @xmath634 these are algebraic set of @xmath635 .
moreover , for a generic choice of @xmath627 , the set @xmath252 is of codimension 1 and @xmath251 is of dimension 0 .
we may replace the conditions of proposition [ prop petit deg top ] by following the idea of the proof of proposition [ prop petit deg top ] , we show that this implies that @xmath638 is of small topological degree on @xmath255 . therefore , we get a zariski open set @xmath370 of mappings asymptotically of small topological degree on attracting sets in @xmath5 .
nevertheless , it is unclear how to find explicit examples of parameters in @xmath370 to ensure that it is not empty .
the last remark around theorem [ theoreme ] is that it is not necessary to be of small topological degree everywhere " on an attracting set in order for it to be non pluripolar .
furthermore , there exists non - pluripolar attracting sets that can not support a current of bidegree ( 1,1 ) of mass 1 with a bounded quasi - potential .
let @xmath0 be an endomorphism of the form @xmath639)=[w^d p\left(\dfrac{z}{w}\right):w^d : t^d+\varepsilon r(z , w)]\ ] ] where @xmath640 is a polynomial of degree @xmath23 and @xmath189 is a homogeneous polynomial of degree @xmath23 .
in particular , it is of the form . on the other hand
, @xmath0 is not asymptotically of small topological degree and the attracting current @xmath28 has a unbounded quasi - potential because @xmath641})= \lbrace l_{[1:0 ] } \rbrace$ ] , hence @xmath642}$ ] is reduced to a point .
nevertheless , there exists a bidisk of the form @xmath15 ; |z|\leq r|w|,|t|<\rho \max ( |z|,|w| ) \rbrace$ ] where @xmath0 is horizontal - like .
moreover , we can adapt the proof of theorem [ theoreme ] to arrange that this horizontal - like map is of small topological degree . for this
, we replace the generic conditions of proposition [ prop petit deg top ] by the following : finally , we adapt the proof of theorem [ th potentiel bourne ] to horizontal - like maps using the canonical potential .
we recall that it is defined as follows ( see @xcite for details ) .
if @xmath69 is a horizontal current , i.e. with support in @xmath648 , its canonical potential is defined by @xmath649 with @xmath650 $ ] .
thus we finally get : * in the light of theorem [ th att cont in hyperplan ] , it would be interesting to find the right hypothesis ensuring non pluripolarity of attracting set in higher codimension .
* does an attracting set of codimension 1 support a current with continuous quasi - potential ? can we prove the convergence of the quasi - potentials ?
dinh ; n. sibony _ dynamics in several complex variables : endomorphisms of projective spaces and polynomial - like mappings . _ holomorphic dynamical systems , 165 - 294 , lecture notes in math . , 1998 , springer , berlin , 2010 .
fornss ; n. sibony ; _ complex dynamics in higher dimension .
modern methods in complex analysis _
( princeton , nj , 1992 ) , 135 - 182 , ann . of math . stud . , 137 , princeton univ . press , princeton , nj , 1995 .
fornss ; n. sibony ; _ dynamics of @xmath3 ( examples ) . laminations and foliations in dynamics , geometry and topology _ ( stony brook , ny , 1998 ) , 47 - 85 , contemp . math .
, 269 , amer . math .
soc . , providence , ri , 2001 . | soit @xmath0 un endomorphisme holomorphe de @xmath1 possdant un ensemble attractif @xmath2 .
dans cet article , nous nous intressons lataille " de @xmath2 , au sens de la gomtrie complexe et de la thorie du pluripotentiel .
nous introduisons un cadre conceptuellement simple permettant dobtenir des ensembles attractifs non algbriques .
nous prouvons quen ajoutant une condition de dimension , ces ensembles supportent un courant positif ferm avec un quasi - potentiel born ( ce qui rpond une question de t.c .
dinh ) .
ils sont donc non pluripolaires . de plus
, nous montrons que les exemples sont abondants dans @xmath3 . |
in this paper we shall study the dynamics of one or more self - gravitating spherical shells of matter subject to a centrally symmetric gravitational field and the application of the ensuing formalism to the semiclassical treatment of hawking radiation .
such field of research was started by the papers by kraus and wilczek @xcite .
the main results we shall report here are the extension of the treatment to one massive shell of matter and also the extension to two or more massive shells of matter which in the time development can also intersect @xcite .
the main appeal of the approach of @xcite is that energy conservation is taken exactly into account and thus the back reaction effects can be computed .
we shall keep the formalism for more that one shell as close as possible to the original formalism of @xcite which can be summarized in words that we shall adopt a painlev gauge .
the extraction of the reduced hamiltonian is often classified as a very complicated procedure .
we shall show that by introducing a `` generating function '' the derivation can be drastically simplified and also by the same token it can be extended to deal not only with massive shells but also with a finite number of massive shells .
the treatment of massless shells is just a particular case . in sec.[reducedsec ]
we shall give the derivation of the reduced action in the case of one shell while proving the independence of the canonical momentum within the painlev class of gauges .
in sec.[eqmotionsec ] we shall discuss the equations of motion . in sec.[analyticsec
] we work out the analytic properties of the conjugate momentum @xmath0 which appears in the reduced action and in sec.[twoshellsec ] we shall briefly describe the extension of the treatment to more than one shell .
an application of the results is given in sec.[exchangesec ] giving a derivation within such a formalism of the dray-t hooft- redmount exchange relations .
an important integrability relation is reported in sec.[integrabilitysec ] which allows to compute the imaginary part of the space integral of the canonical momenta for two shells . in sec.[modesec ] we revisit the role of the imaginary part of the canonical momentum in determining the bogoliubov coefficients and thus the most important features of hawking radiation . in sec.[conclusionsec ] we give some concluding remarks .
as usual we write the metric for a spherically symmetric configuration in the adm form @xmath1 were following @xcite we shall choose the functions @xmath2 as continuous functions of the coordinates .
we shall work on a finite region of space time @xmath3 .
on the two initial and final surfaces we give the intrinsic metric by specifying @xmath4 and @xmath5 and similarly @xmath6 and @xmath7 .
the complete action in hamiltonian form , boundary terms included is @xcite @xmath8 where @xmath9 @xmath10 is the shell position and @xmath11 its canonical conjugate momentum . @xmath12 and @xmath13 are the constraints @xmath14 @xmath15 action ( [ completeaction ] ) is immediately generalized to a finite number of shells .
the shell action as given by eqs.([shellaction],[constraintr],[constraintt ] ) refers to a dust shell even though generalizations to more complicated equations of state have been considered @xcite .
the boundary terms are those given in the paper by hawking and hunter @xcite and will play a very important role in the following .
the function @xmath16 @xcite @xmath17 has the remarkable property of generating the conjugate momenta as solutions of the constraints as follows @xmath18 @xmath19 where @xmath20 is a mass which is constant in @xmath21 except at the shell position @xmath10 and that we shall denote by @xmath22 for @xmath21 above all the shell positions and by @xmath23 for @xmath21 below all the shell positions .
we shall adopt a painlev gauge defined by @xmath24 everywhere .
with regard to the remaining freedom in the choice of the gauge we shall choose @xmath25 except for a deformation region near the shell positions .
such deformation is unavoidable because the constraints impose a discontinuity in the derivative @xmath26 at @xmath27 as we shall see in the following . in the painlev gauges @xmath16
becomes @xmath28 where @xmath29 and @xmath30 where we exploited the freedom of adding to ( [ fgeneral ] ) a total derivative , thus gaining for @xmath16 the useful property of vanishing wherever @xmath31 . for @xmath24 eqs.([pilfromf],[pirfromf ] ) become @xmath32}{w}.\ ] ] with regard to @xmath33 one can choose several gauges , within the painlev family .
inner gauge ] inner gauge ] generic gauge ] typical are the `` outer gauge '' characterized by @xmath34 for @xmath35 and shown in fig.[outer ] and the inner gauge characterized by @xmath34 for @xmath36 shown in fig.[inner ] , but there are also more general choices as shown in fig.[generic ] . we stress that such deformation @xmath37 is not related to the thickness of the shell which is always zero . using the solutions ( [ pilfromf],[pirfromf ] ) of the constraints one obtains in the outer gauge the following reduced @xcite action where only the shell position @xmath10 appears as degree of freedom @xmath38 the @xmath0
is easily computed @xmath39 where @xmath40 and similarly for @xmath41 . in deriving eq.([pcgeneral ] ) we used the consequences of the constraints ( [ constraintr],[constraintt ] ) @xmath42 in the outer gauge we find @xmath43 with @xmath11 given implicitly by @xmath44 this is the result obtained by friedman , louko , winters - hilt @xcite through a limit procedure in which the support of the deformation function goes to zero but actually it is independent of the deformation . for working out the dynamics of one shell
the limit procedure is all right , but in dealing with two or more shell it is important to have a deformation on a finite range because otherwise a limit procedure would give a result dependent on the order in which the two limits , deformation and @xmath45 , are taken .
similarly one can compute using the general formula the reduced canonical momentum of the system in the inner gauge @xmath46 and @xmath11 , again determined by the discontinuity equation ( [ disceq ] ) , is given now by the implicit equation @xmath47 the two reduced canonical momenta @xmath48 appear to be completely different but they can be proven to be the same function of @xmath10 .
one can consider also the more general gauges depicted in fig.3 and it can be proven @xcite that @xmath0 is always the same .
however in inner gauge a term @xmath49 appears in the reduced action while in the more general gauges both a term @xmath50 and a term @xmath51 appear in the action .
the boundary term given in eq.([completeaction ] ) is equivalent to @xmath52
in deriving the equation of motion from action ( [ outerreducedaction ] ) one can consider @xmath53 , the interior mass , as a datum of the problem and vary @xmath22 to obtain @xmath54 which using the expression of @xmath0 and the relation between @xmath55 and @xmath56 imposed by the gravitational equations @xcite can be written as @xmath57 alternatively one can consider @xmath22 , the total energy as a datum of the problem and vary @xmath58 .
the calculation is far more complicated due to the presence of @xmath50 in the action ( [ outerreducedaction ] ) , but using the equations for the gravitational field one reaches @xcite the same equation of motion ( [ 1sheqmotion ] ) .
in addition a consequence of the gravitational equations of motion is the constancy in time of @xmath58 and @xmath59 .
we saw that in the outer gauge @xmath0 is given by ( [ pc],[fundamentalh ] ) .
the solution of eq.([fundamentalh ] ) for @xmath11 is @xmath60 where @xmath61 if we want @xmath11 to describe an outgoing shell we must choose the plus sign in front of the square root .
moreover the shell reaches @xmath62 if and only if @xmath63 as expected .
the logarithm in @xmath0 , eq.([pc ] ) , has branch points at zero and infinity and thus we must investigate for which values or @xmath10 such values are reached . at @xmath64 , @xmath11 has a simple pole with positive residue .
thus the numerator in the argument of the logarithm in @xmath0 goes to zero and below @xmath65 it becomes @xmath66 where here @xmath67 is the absolute value of the square root .
expression ( [ contnumerator ] ) is negative irrespective of the sign of @xmath11 and stays so for @xmath68 because @xmath11 is no longer singular . as a consequence @xmath0 below @xmath65 acquires the imaginary part @xmath69 . below @xmath70 the denominator of the argument of the logarithm in eq.([pc ] )
becomes negative so that the argument of the logarithm reverts to positive values .
thus the so called classically forbidden region where @xmath0 becomes complex is @xmath71 independent of @xmath72 and of the deformation @xmath37 and the integral of the imaginary part of @xmath0 for any deformation @xmath37 and for any mass @xmath72 of the shell is @xmath73 which is the original result derived in @xcite in the zero mass case .
parikh and wilczek @xcite gave to @xmath74 the interpretation of the tunneling probability for the emission of a quantum of energy @xmath75 .
criticism and alternative proposals for the emission probability like @xmath76 @xmath77 followed @xcite . here instead we shall discuss the role of eq.([pc ] ) in the framework of mode analysis .
this will be done in sec.([modesec ] ) .
two shell dynamics ] now we have three characteristic masses , @xmath23 , @xmath22 and an intermediate mass @xmath78 , see fig.4 , which can change only if the two shells cross . working as before with @xmath79 considered as a datum of the problem
we reach the reduced action @xcite @xmath80 where @xmath81 and @xmath82 is given by eq.([pc ] ) with @xmath23 replaced by @xmath78 .
the novelty is that now even in the outer gauge the time derivative of @xmath22 intervenes in addition to @xmath83 and @xmath84 and @xmath85 depend both on @xmath86 and @xmath87 .
we can vary @xmath86 , @xmath87 , @xmath22 and @xmath78 independently obtaining the correct equations of motion @xcite . as expected
one finds that the exterior shell moves irrespective of the dynamics which develops at lower values of @xmath21 until a crossing occurs .
in case of crossing of the two shells from the equations of sec.([twoshellsec ] ) we can obtain relations between @xmath88 and @xmath89 being @xmath90 the the shell position at the crossing and @xmath89 the intermediate mass after the crossing . during the crossing the masses of the shells can change ,
provided they satisfy a relation analogous to the energy - momentum conservation is special relativity @xmath91 this is an outcome of the constraints .
it is not possible to predict the final masses of the two shells as it depends on the details of the interaction which has to be specified .
a relatively simple case is when the masses are unchanged during the crossing ( transparent crossing ) @xcite and the simplest case is the crossing of two massless shells which remain massless , thus re - obtaining the well known dray-t hooft - redmount relations @xcite which we report here below @xmath92 being @xmath90 the crossing radius .
the exchange diagram ]
the space - part of the on - shell action for two massive shells of matter is given from eq.([twoshellreducedaction ] ) by @xmath94 as on the equations of motion @xmath95 .
it is possible to prove a theorem analogous to the one found in the books of whittaker and arnold @xcite i.e. that in presence of a constant of motion , in addition to the hamiltonian , the form @xmath96 is closed even though the proof is somewhat different @xcite .
the intermediate mass @xmath78 plays the role of the additional constant of motion .
the above result allows to deform the integration path as to bring @xmath97 immediately below or above @xmath87 .
this is allowed by the absence of discontinuities in our scheme .
but , in words , two coaleshed shell have the same properties of a single shell with energy @xmath98 .
the final result is that @xmath99 the result ( [ intpc1pc2 ] ) holds also for in the case when at the crossing the two shells can change their mass .
for details we refer to @xcite .
the original way to extract information on the spectrum of the radiation is mode analysis @xcite . in the massless case @xmath100
can be computed exactly and its expansion to first order in @xmath101 is @xmath102 given the semiclassical mode @xmath103 we can perform the analysis of the mode regular at the horizon in terms of the above mode .
this analysis can be performed either by scalar product @xcite i.e. space integration where one has to keep into account that the background metric is painlev or by time fourier analysis . in this way one
obtains the well known formulas for the bogoliubov coefficients .
the regular outgoing modes near the horizon are given by @xmath104 computing the scalar product , and taking into account that the background metric is the painlev metric , we have @xmath105 with @xmath106 and the integration region is outside the horizon . keeping only the most singular terms at the horizon eq.([scalarproduct ] )
reduces , with @xmath107 , to @xmath108 we can compute such scalar products at @xmath109 obtaining the standard hawking integral @xmath110 which gives the dominant contribution for large @xmath111 . the coefficient @xmath112 is obtained by changing in ( [ alphakomega ] ) @xmath75 into @xmath113 .
alternatively we can extract the bogoliubov coefficients by performing a time fourier transform i.e. @xmath114 which is , as it should be , a result independent of @xmath21 ( i.e. @xmath115 ) and reproduces eq.([alphakomega ] ) .
the derivation is valid to first order in @xmath75 which is the realm of validity of the external field approximation . for finite @xmath75
it is problematic to perform a space integration on modes , because we have not a well defined background metric .
thus kraus and wilczek @xcite followed the time fourier analysis method . in order to do
so one has to construct the non - perturbative modes regular at the horizon .
this is not a completely trivial task .
the regular modes are given by @xmath116 where @xmath117 is the on shell action computed with the following boundary conditions : 1 ) at time @xmath118 the shell position is @xmath10 , outside the horizon .
2 ) at time @xmath109 the conjugate momentum is @xmath111 , i.e. @xmath119 . due to these boundary conditions @xmath120 , @xmath22 and as a consequence @xmath0 depend on @xmath121 even if along each trajectory @xmath22 is always a constant of motion . using the saddle point approximation @xcite one obtains that the absolute value of the bogoliubov coefficient @xmath122 is given by @xmath123 computed for @xmath124 .
what is important here is that only the space part of the action appears , as the time part @xmath125 cancels with @xmath126 at the saddle point .
@xmath120 is given by the condition @xmath127 which for @xmath128 is solved by @xmath129 and thus there is no imaginary contribution to the integral from the gap @xmath130 as we discussed in sect.([analyticsec ] ) .
the bogoliubov coefficient @xmath112 is obtained by changing in ( [ kvkintegral ] ) @xmath75 in @xmath113 ( always @xmath131 ) . then we have @xmath132 the value of the saddle point @xmath118 , now complex , is given by @xmath133 and the r.h.s .
term of eq.([kvkintegral ] ) becomes @xmath134 where the last passage is due to eq.([integratedimpart ] ) for which we gave a general proof within the family of painlev gauges even in the massive case .
an explicit derivation of the result ( [ correctedbeta ] ) has been given in @xcite by working out the late time expansion of the time development of the action .
the main point is that in the mode analysis what comes in is not the total action of a model particle crossing the horizon , but only the `` space part '' of it .
the analysis and the results depend on the interpretation of the expression @xmath135 as modes of the field dressed by the gravitational interaction with non vanishing quanta of energy .
the result eq.([intpc1pc2 ] ) proven for the emission of two shells @xcite was interpreted @xcite as the absence of correlations among quanta in the emitted radiation .
it would be of interest to give a similar `` mode interpretation '' of the result ( [ intpc1pc2 ] ) derived for the emission of two shells which during the time evolution can also interact @xcite . to this end
one should compute the two shell modes which are regular at the horizon , and perform a time fourier analysis of them .
this has not yet been accomplished .
in this paper we gave a general treatment of the dynamics of one or more self- gravitating spherical shell of matter in a spherical gravitational field .
we extended the treatment of @xcite to more than one shell @xcite proving on the way the universality of the reduced conjugate momentum within the family of painlev gauges @xcite .
this allows to give a general derivation of the exchange relations and also , exploiting an integrability result , to extend the result on the imaginary part of the space part of the action to more than one shell .
instead of following the tunneling picture here we have revisited the treatment which follows by interpreting the exponential of the action , properly subtracted , as the dressed modes of the hawking radiation .
the main point here is that according such a mode interpretation only the space part of the action and in particular its imaginary part comes in determining the bogoliubov coefficient and thus in determining the hawking spectrum corrected for the back reaction effects .
9 n.d . birrel and p.c.w .
davies , _ quantum field theory in curved space _ cambridge monographs on mathematical physics , ( cambridge university press , cambridge u.k . 1982 ) ; r. wald , _ quantum field theory in curved spacetime and black- hole thermodynamics _ , ( the university of chicago press , chicago usa 1994 ) | we consider the dynamics of one or more self gravitating shells of matter in a centrally symmetric gravitational field in the painlev family of gauges . we give the reduced hamiltonian for two intersecting shells , both massless and massive . such a formulation is applied to the computation of the semiclassical action of two intersecting shells .
the relation of the imaginary part of the space - part of the action to the computation of the bogoliubov coefficients is revisited . |
there exist two distinct types of supernova explosions : one is type ii supernovae ( sne ii ) , which are the core collapse - induced explosions of short - lived massive stars ( @xmath4 ) and produce more o and mg relative to fe ( i.e. , [ o / fe ] @xmath5 ) , and the other is type ia supernovae ( sne ia ) , which are the thermonuclear explosions of accreting white dwarfs ( wds ) in close binaries and produce mostly fe and little o. the exact companion stars of the wds have not been identified but must be relatively long - lived low mass stars .
the role of these two types of supernovae in the chemical evolution of galaxies can be seen in the abundance ratios of stars with different metallicities , most notably in the [ o / fe]-[fe / h ] relation .
metal poor stars with [ fe / h ] @xmath0 have [ o / fe ] @xmath6 on the average ( @xcite ; @xcite ) , while disk stars with [ fe / h ] @xmath7 show a decrease in [ o / fe ] with increasing metallicity ( @xcite ; @xcite ; @xcite ) . such an evolutionary change in [ o / fe ] against [ fe / h ]
has been explained with the early heavy element production by sne ii and the delayed enrichment of fe by sne ia ( @xcite ) .
conversely , chemical evolution models can constrain the nature of still uncertain progenitor systems of sne ia ; for example , yoshii , tsujimoto , & nomoto ( 1996 ) estimated the lifetime of sn ia progenitors as long as @xmath8 gyr .
sne ia have been discovered up to @xmath9 by supernova cosmology project ( @xcite ) and high - z supernova search team ( @xcite ) .
they have given the sn ia rate at @xmath10 ( @xcite ) but will provide the sn ia rate history over @xmath11 . with the next generation space telescope , both sne ia and ii will be observed through @xmath12 . in theoretical approach , the cosmic sn ia rate as a function of redshift has been constructed for a cosmic star formation rate ( sfr ) ( @xcite ; @xcite ; @xcite ; @xcite ) . the comparison between the model prediction and observations can constrain the lifetime of the progenitor systems of sne ia .
the progenitors of the majority of sne ia are most likely the chandrasekhar ( ch ) mass wds ( e.g. , @xcite for a recent review ) , although the sub - ch mass models might correspond to some peculiar sub - luminous sne ia .
the early time spectra of the majority of sne ia are in excellent agreement with the synthetic spectra of the ch mass models , while the spectra of the sub - ch mass models are too blue to be compatible with observations ( @xcite ; @xcite ) . for the evolution of accreting wds toward the ch mass ,
two scenarios have been proposed : one is a double degenerate ( dd ) scenario , i.e. , merging of double c+o wds with a combined mass surpassing the ch mass limit ( @xcite ; @xcite ) , and the other is a single degenerate ( sd ) scenario , i.e. , accretion of hydrogen - rich matter via mass transfer from a binary companion ( e.g. , @xcite for a review ) .
the issue of dd vs. sd is still debated ( e.g. , @xcite for a review ) , although theoretical modeling has indicated that the merging of wds does not make typical sne ia ( @xcite , 1998 ) . for the sd scenario ,
a new evolutionary model has been proposed by hachisu , kato , & nomoto ( 1996 , 1998 ; hereafter hkn96 and hkn98 , respectively ) and hachisu & kato ( 1998 ; hereafter hk98 ) .
hkn96 have shown that if the accretion rate exceeds a certain limit , the wd blows a strong wind and burns hydrogen steadily to increase the wd mass ( see section 2 ) . hkn98
have further invoked the effect of stripping - off of the envelope from the companion by the strong wind , and shown that the wd can reach the ch mass for much wider binary parameter space than found by hkn96 , li & van den heuvel ( 1997 ) , and yungelson & livio ( 1998 ) ; the allowed parameter space may be large enough to account for the sn ia frequency .
moreover , hk98 have found an important metallicity effect ; if the iron abundance of the accreted matter is as low as [ fe / h ] @xmath0 , the wd wind is too weak to increase the wd mass through the ch mass . in the present _ letter _
, we apply the above two scenarios to the chemical evolution models , and compare the cases with and without the metallicity effect on sne ia .
we have found that the model for sd scenario with the metallicity effect is significantly better to reproduce the evolutionary change in [ o / fe ] and other properties ( section 3 ) . using the metallicity dependent sn ia rate
, we make a prediction for the cosmic supernova rate history ( section 4 ) . in section 5
we discuss other implications on the galactic halo objects , damped ly@xmath3 ( dla ) systems and cosmology .
our sd scenario has two progenitor systems : one is a red - giant ( rg ) companion with the initial mass of @xmath13 and the orbital period of tens to hundreds days ( hkn96 ; hkn98 ) . the other is a near main - sequence ( ms ) companion with the initial mass of @xmath14 and the period of several tenths of a day to several days ( @xcite ; hkn98 ) . in these sd scenarios ,
optically thick winds from the mass accreting wd play an essential role in stabilizing the mass transfer and escaping from forming a common envelope .
the optically thick winds are driven by a strong peak of opal opacity at @xmath15 ( e.g. , @xcite ) .
since the peak is due to iron lines , the optically thick winds depend strongly on the metallicity ( hk98 ) .
figure 1 shows the metallicity dependence of the optically thick winds .
the strong winds are possible only for the region above the dashed line .
the term `` weak '' implies that the wind velocity at the photosphere does not exceed the escape velocity there , that is , it can not blow the accreted matter off the wd . for the metallicity as small as @xmath16 ,
the opacity peak at @xmath17 is very weak , being smaller than the peak of helium lines at @xmath18 .
then , the wind is driven by the helium line peak rather than the iron line peak , which we call `` he wind '' instead of `` fe wind '' . since only the initial wd mass of @xmath19 can produce an sn ia ( @xcite ) , sn ia events occur only for the progenitors with [ fe / h ] @xmath20 , which is adopted in our chemical evolution model .
figure 2 shows the sn ia regions in the diagram of the initial orbital period vs. the initial mass of the companion star for the initial wd mass of @xmath21 ( see hk98 for other @xmath22 ) . in these regions ,
the accretion from the companion star increases the wd mass successfully through the occurrence of sn ia .
the dashed line shows the case of solar abundance ( @xmath23 ) , while the solid line shows the much lower metallicity case of @xmath24 .
the size of these regions clearly demonstrate the metallicity effect , i.e. , sn ia regions are much smaller for smaller metallicity .
the initial mass ranges of the companion stars for @xmath24 are @xmath25 for the wd+rg system and @xmath26 for the wd+ms system .
we use the chemical evolution model which allows the material infall from outside the disk region . for the infall rate , we adopt a formula which is proportional to @xmath27 $ ] with a infall timescale of @xmath28 gyr ( yoshii et al .
the galactic age is assumed to be 15 gyr .
the sfr is assumed to be proportional to the gas fraction with a constant rate coefficient of @xmath29 gyr@xmath30 .
for the initial mass function ( imf ) , we assume a power - law mass spectrum with a salpeter slope of @xmath31 in the range of @xmath32 ( @xcite ) . we take the nucleosynthesis yields of sne ia and ii from tsujimoto et al .
( 1995 ; see nomoto et al .
1997b , c for details ) and the metallicity dependent main - sequence lifetime from kodama ( 1997 ) . for the sd scenario ,
the lifetime of sne ia is determined from the main - sequence lifetime of the companion star .
we adopt the initial mass ranges of the binary companion stars obtained in section 2 .
the distribution function of the initial mass of the companions is taken from the mass ratio distribution in binaries ( @xcite ) , which is approximated by a power - law mass spectrum with a slope @xmath33 .
the fraction of primary stars of @xmath34 which eventually produce sne ia is set to be @xmath35 for both the wd+ms and the wd+rg systems , adjusted to reproduce the chemical evolution in the solar neighborhood . for the dd scenario
we adopt the distribution function of the lifetime of sne ia by tutukov & yungelson ( 1994 ) , majority of which is @xmath36 gyr .
figure 3 shows the evolutionary change in [ o / fe ] for three sn ia models .
the dotted line is for the dd scenario .
the other lines are for our sd scenario with ( solid line ) and without ( dashed line ) the metallicity effect on sne ia .
the results are as follows .
* in the dd scenario the lifetime of the majority of sne ia is shorter than @xmath37 gyr .
then the decrease in [ o / fe ] starts at [ fe / h ] @xmath38 , which is too early compared with the observed decrease in [ o / fe ] starting at [ fe / h ] @xmath39 .
* for the sd scenario with no metallicity effect , the companion star with @xmath40 evolves off the main - sequence to give rise to sne ia at the age of @xmath41 gyr .
the resultant decrease in [ o / fe ] starts too early to be compatible with the observations .
* for the metallicity dependent sd scenario , sne ia occur at [ fe / h ] @xmath7 , which naturally reproduce the observed break in [ o / fe ] at [ fe / h ] @xmath39 .
we also perform the kolmogorov - smirnov ( ks ) test to estimate how well the predicted abundance patterns agree with the observed one ( @xcite ) , and tabulate the resultant probabilities in table 1 .
the ks test is applied to the abundance distribution function in the range of @xmath42 } \leq 0.45 $ ] and to the evolutionary behavior of [ o / fe ] in the range of @xmath43 } \leq 0.02 $ ] .
the predicted abundance distribution functions are identical with the observational data with more than @xmath44 probability for both the dd and sd scenarios . for our sn ia model with the metallicity effect , the calculated [ o / fe ] against [ fe / h ] fits well to the observations with @xmath45 probability .
but when we neglect the metallicity effect , the probability decreases to @xmath46 . for the dd scenario
the calculated [ o / fe ] is rejected with @xmath47 probability .
we apply our sn ia progenitor scenario to predict the cosmic supernova rate history and the cosmic chemical evolution corresponding to the observed cosmic sfr ( @xcite ; @xcite ) .
the photometric evolution is calculated with the spectral synthesis population database taken from kodama ( 1997 ) .
we adopt @xmath48 km s@xmath30 mpc@xmath30 , @xmath49 , @xmath50 , and the redshift at the formation epoch of galaxies @xmath51 .
we use the initial comoving density of gas @xmath52 ( @xcite ) .
figure 4 shows the cosmic supernova rate per @xmath53 per century ( snu ) .
the long dashed line is for sne ii and the other lines for sne ia with ( solid line ) and without ( dashed line ) the metallicity effect .
if we do not include the metallicity effect , the sn ia rate is almost flat from the present to higher redshift , and decreases toward the formation epoch of galaxies . if we include the metallicity effect , the sn ia rate drops at @xmath54 , where the iron abundance of the gas in the universe is too low ( i.e. , [ fe / h ] @xmath0 ) for the progenitors of sne ia to make explosions .
the redshift where the sn ia rate drops is determined by the speed of the chemical enrichment , which depends on the effect of dust extinction on the cosmic sfr ( @xcite ) , cosmology , galaxy formation epoch and the initial gas density . taking into account these uncertainties , the break in the sn ia rate occurs at @xmath55 .
we should note that galaxies being responsible for the global sfr have different heavy - element enrichment timescale , some of which are starburst galaxies and ellipticals achieving [ fe / h ] @xmath39 in much shorter timescale than in figure 4 . in such galaxies sne ia can occur even at @xmath2 .
we introduce a metallicity dependence of the sn ia rate in the galactic and cosmic chemical evolution models . in our scenario involving a strong wind from wds , little sne
ia occur at [ fe / h ] @xmath0 .
our model successfully reproduces the observed chemical evolution in the solar neighborhood .
we make the following predictions which can test this metallicity effect .
1 ) sne ia are not found in the low - iron abundance environments such as dwarf galaxies and the outskirts of spirals .
2 ) the cosmic sn ia rate drops at @xmath1 due to the low - iron abundance , which can be observed with the next generation space telescope . at @xmath2
, sne ia can be found only in the environments where the timescale of metal enrichment is sufficiently short as in starburst galaxies and ellipticals .
\1 ) it imposes a limit on the metallicity range of the sn ia progenitors .
this would imply that `` evolution effects '' are relatively small for the use of high redshift sne ia to determine the cosmological parameters .
\2 ) microlensing experiments ( @xcite ) suggest the wd - dominated galactic halo .
the existence of so many wds results in too much iron enrichment from sne ia ( @xcite )
. however , wds of halo populations are poor producers of sne ia , so that the iron abundance of the halo may not put the strong constraint on the contribution of wds .
note , however , that there exist several other arguments against the wd - dominated halo ( e.g. , @xcite ; @xcite ) .
\3 ) the galactic halo is a low - metallicity system with [ fe / h ] @xmath0 and have an abundance pattern of genuine sn ii origin , i.e. , the overabundances of @xmath3-elements relative to fe as [ @xmath3/fe ] @xmath5 ( e.g. , @xcite ) .
however there exist age differences of several gyrs among the clusters ( @xcite ) as well as field stars ( @xcite ) .
since the shortest lifetime of sne ia is @xmath41 gyr for the ms+wd close binary systems , sn ia contamination would be seen in [ @xmath3/fe ] if there were no metallicity effect on sne ia .
this apparent discrepancy between the age difference and the high [ @xmath3/fe ] can be resolved by the low - metallicity inhibition of sne ia .
\4 ) similar interpretation holds also for dla systems .
the dla systems observed at @xmath56 have [ fe / h]@xmath57 to @xmath58 and indicate [ @xmath3/fe ] @xmath5 ( @xcite ) .
( 1996 ) suggested there may exist the age - metallicity relation in dla systems , which implies that dla systems grown through a common chemical history spanning over several gyrs .
if so , the sn ii - like abundance pattern in dla systems needs the introduction of the metallicity dependent sn ia rate to avoid the contamination of sn ia products .
this work has been supported in part by the grant - in - aid for scientific research ( 05242102 , 06233101 , 08640336 , 08640321 , 09640325 ) and coe research ( 07ce2002 ) of the ministry of education , science , culture , and sports in japan .
we would like to thank tadayuki kodama for providing us with the database of simple stellar population spectra , and david branch for many useful comments to improve the paper . | we introduce a metallicity dependence of type ia supernova ( sn ia ) rate into the galactic and cosmic chemical evolution models . in our sn
ia progenitor scenario , the accreting white dwarf ( wd ) blows a strong wind to reach the chandrasekhar mass limit .
if the iron abundance of the progenitors is as low as [ fe / h ] @xmath0 , then the wind is too weak for sne ia to occur .
our model successfully reproduces the observed chemical evolution in the solar neighborhood .
we make the following predictions which can test this metallicity effect : 1 ) sne ia are not found in the low - iron abundance environments such as dwarf galaxies and the outskirts of spirals .
2 ) the cosmic sn ia rate drops at @xmath1 due to the low - iron abundance , which can be observed with the next generation space telescope . at @xmath2
, sne ia can be found only in the environments where the timescale of metal enrichment is sufficiently short as in starburst galaxies and ellipticals .
the low - metallicity inhibition of sne ia can shed new light on the following issues : 1 ) the limited metallicity range of the sn ia progenitors would imply that `` evolution effects '' are relatively small for the use of high redshift sne ia to determine the cosmological parameters .
2 ) wds of halo populations are poor producers of sne ia , so that the wd contribution to the halo mass is not constrained from the iron abundance in the halo .
3 ) the abundance patterns of globular clusters and field stars in the galactic halo lack of sn ia signatures in spite of their age difference of several gyrs , which can be explained by the low - metallicity inhibition of sne ia .
4 ) it could also explain why the sn ia contamination is not seen in the damped ly@xmath3 systems for over a wide range of redshift . |
big bang nucleosynthesis ( bbn ) model @xcite successfully explains primordial light element abundances inferred from astronomical observations ( e.g. @xcite ) if the cosmological baryon density determined from the power spectrum of cosmic microwave background ( cmb ) radiation measured with the wilkinson microwave anisotropy probe ( wmap ) @xcite or planck @xcite is adopted .
an apparent discrepancy , however , exists between observational and theoretical @xmath6li abundances .
spectroscopic observations of metal - poor stars ( mpss ) indicate a plateau for the abundance ratio , @xmath6li / h@xmath14 , with small error bars as a function of metallicity for [ fe / h]@xmath15@xmath16 , where @xmath17 is the number density of element ( @xmath18=a and b ) , and the subscript @xmath19 indicates the solar value . ] in the galaxy @xcite and @xmath20 centauri accreted by the galaxy @xcite.li / h)@xmath21 @xcite , while the li abundance in halo dwarf stars is log(@xmath6li / h)@xmath22 @xcite . however , these abundances are consistent with each other within the uncertainties for the moment ( see fig . 3 of ref .
the possible systematic difference in the li abundances in gcs and galactic halo should be studied further in future . ]
the plateau abundance is @xmath23 times smaller than that predicted in standard bbn ( sbbn ) model ( e.g. , @xmath6li / h=@xmath24 @xcite ; see ref .
@xcite for theoretical light element abundances for the baryon density from planck @xcite ) .
recent observations also indicate a break of this plateau shown as small li abundances with large dispersion at lower metallicities of [ fe / h]@xmath25 @xcite .
therefore , it seems that we need a mechanism for a metallicity - independent depletion from the primordial abundance to the plateau abundance and also another for a metallicity - dependent depletion from the plateau abundance . in this paper , we focus on the former universal depletion by cosmological processes .
this li problem ( see ref .
@xcite for a review ) shows that some physical processes have reduced the primordial li abundance in some epoch during or after bbn .
standard stellar model suggests very small depletions of li isotopes in surfaces of mpss @xcite .
the @xmath6li / h abundances of mpss observed today are then approximately interstellar abundances when the mpss formed .
nonstandard processes such as the rotationally induced mixing @xcite , and the turbulent mixing @xcite have been suggested to reduce the @xmath6li abundance in stellar atmospheres . in the former model ,
a large depletion factor does not realize simultaneously with a small dispersion in stellar li abundances after the depletion .
the depletion factor is then constrained to be small . in the latter model , a depletion of a factor of @xmath26 @xcite is predicted although it is still unclear if this mechanism can deplete li abundances of all mpss rather uniformly .
nonstandard bbn , on the other hand , may be responsible for the li problem at least partially .
we note that @xmath6be is produced more than @xmath6li in sbbn model with the planck baryon density .
the @xmath6be nuclei are then converted to @xmath6li nuclei via recombination with electron followed by the electron capture , i.e. , @xmath6be + e@xmath27li + @xmath28 .
therefore , the li problem is alleviated if some exotic processes could destroy @xmath6be .
one of solutions to the li problem is an injection of nonthermal photon with energy of @xmath29 mev which can destroy @xmath6be but not deuterium ( d ) as calculated in ref .
if a long - lived exotic particle radiatively decays after bbn , nonthermal photons can disintegrate background thermal nuclei , and light element abundances change @xcite .
if the energy of the photon emitted at the decay is much larger than @xmath30 mev , all of light nuclei are disintegrated by nonthermal photons @xcite . in this case , therefore , the li problem can not be solved ( e.g. , @xcite ) .
therefore , the energy of emitted photon for the @xmath6be destruction is limited to a narrow range @xcite .
a similar @xmath6be destruction would occur if a long - lived sterile neutrino decays into energetic electron and positron .
we then study cosmological effects of this decay channel in this paper . from the theoretical point of view of the extended minimal standard model ( msm ) of particle physics , right - handed neutrinos introduced as sterile neutrinos provide an elegant mechanism for the generation of tiny active neutrino masses ; so - called canonical seesaw mechanism @xcite .
if their masses are so heavy ( more than @xmath31 ) , they also explain the origin of the baryon asymmetry of the universe ; so - called leptogenesis scenario @xcite .
even if the sterile neutrinos have masses below electroweak ( ew ) scale , however , there exist other phenomenological effects without lacking the success of the seesaw mechanism .
several possibilities have been investigated regarding a detectability of the sterile neutrinos , and the case in which sterile neutrinos are lighter than light mesons ( e.g. , pion or kaon ) has been studied especially in detail @xcite ( for recent study , see @xcite and references therein ) .
the allowed sterile neutrino masses are smaller than the pion mass @xmath32 140 mev , which are not consistent with recent results of neutrino oscillation experiments , unless one assumes that their lifetimes are longer than @xmath33 s. in addition , another constraint has been derived from a study of bbn by a comparison between theoretical and observational abundances of @xmath34he .
the upper limit on the lifetime of @xmath33 s was derived when a relic abundance of sterile neutrino is fixed as given in ref .
however , this constraint depends on the relic abundance . in this paper
, we take into account the possibility that the abundance is smaller than the simple estimate @xcite . in this case a longer lifetime of the sterile neutrino is allowed . in this paper
, we comprehensively investigate cosmological effects of a long - lived sterile neutrino with mass @xmath35 mev . in sec .
[ sec2 ] , we assume a decay of a sterile neutrino in the early universe , and describe our calculation method , and formulations of ( 1 ) the spectra of electrons and positrons generated at the decay , ( 2 ) those of primary photons induced by the energetic electrons and positrons , and ( 3 ) the nonthermal nucleosynthesis triggered by the energetic photons . in sec .
[ sec3 ] , revised cross sections for photodisintegration of @xmath6be and @xmath6li are described . in sec .
[ sec4 ] , effects of the decaying sterile neutrino on the cosmic thermal history , the effective neutrino number , and the baryon - to - photon ratios are formulated . in sec .
[ sec5 ] , observational constraints on primordial light element abundances , the effective neutrino number , and the baryon - to - photon ratio adopted in this paper are summarized . in sec .
[ sec6 ] , we show calculated energy spectra of electrons and positrons emitted at the decay , energy spectra of photons produced via the inverse compton scattering of background photons by the electron and positron , and photon injection spectra resulting from electromagnetic cascade showers .
time evolutions of light element abundances , the baryon - to - photon ratio , and thermal and nonthermal neutrino energy densities are then consistently calculated with nonthermal photodisintegrations of nuclei taken into account .
an impact of revised cross sections of @xmath6be and @xmath6li photodisintegrations is also shown . in sec .
[ sec7 ] , we discuss the relic abundance of the sterile neutrino before its decay .
we also comment on a possible dilution of the sterile neutrino in the early universe , effects of the sterile neutrino mixing to active neutrinos of different flavors , and experimental constraints from the pion decay and the supernova luminosity . in sec .
[ sec8 ] , this study is summarized . in appendix
[ app1 ] , extensive formulae of the sterile neutrino decay are derived .
we adopt natural units of @xmath36 , where @xmath37 is the reduced planck constant , @xmath38 is the speed of light , and @xmath39 is the boltzmann constant . we also adopt notation of @xmath40(@xmath41 , @xmath42)@xmath43 for a reaction @xmath44 .
we simply include one right - handed ( sterile ) neutrino , @xmath1 , in the msm and assume that it has a mass of @xmath45(10 ) mev and a rather long lifetime @xmath46
s. we consider effects of the decays of @xmath47 sterile neutrino on cosmological quantities , especially the @xmath6li number abundance relative to hydrogen , i.e. , @xmath48/h .
we perform a bbn calculation .
kawano s bbn code @xcite is utilized with default settings of the time steps in order to make the thermal nucleosynthesis calculation part as simple as possible . to calculated results , we added sarkar s correction for @xmath34he abundances from explicit integration of weak rates , smaller time steps , coulomb , radiative , and finite temperature corrections , and the correction for finite nucleon mass @xcite .
reaction rates for light nuclei ( @xmath49 ) are updated with recommendations by jina reaclib database v1.0 @xcite .
the neutron lifetime is set to be @xmath50 s from the weighted average value of the particle data group @xcite .
we note that after the improved measurements @xcite , a few earlier measurements have been reanalyzed , and updated lifetimes are significantly shorter than the previous ones @xcite .
we adopt cosmological parameters reported from the analysis of the planck @xcite .
central values for the base @xmath51cdm model ( planck+wp+highl+bao ) determined from the planck 2013 data are taken : @xmath52 km s@xmath53 mpc@xmath53 , @xmath54 , @xmath55 , and @xmath56 with @xmath57(100 km s@xmath53 mpc@xmath53 ) .
the baryon - to - photon ratio is calculated using most recent values of physical constants as follows .
the present number density of cmb is given by @xmath58 where @xmath59 is the riemann zeta function , and @xmath60 is the present cmb temperature .
the present energy density of baryon is related to cosmological parameters as @xmath61 where @xmath62 is the gravitational constant .
the present average mass per baryon is given @xcite by @xmath63\right\ } m_{\rm h}\nonumber \\ & = & \left(1 -0.007119 y_{\rm p } \right ) m_{\rm h } , \label{m_b0}\end{aligned}\ ] ] where @xmath64 is the present number density of baryon , @xmath65 and @xmath66 kg are the atomic masses of @xmath67h and @xmath34he , respectively @xcite , and @xmath68 is the cosmological mass fraction of @xmath34he . in the above equation ,
contributions from small abundances of d , @xmath13he , @xmath6li , and heavier nuclides have been neglected . using the above three equations ,
the relation between the baryon - to - photon ratio and the baryon density parameter is given @xcite by @xmath69 \frac{m_{\rm h}}{\rho_{{\rm
b}0 } } \nonumber\\ & = & \frac{1}{n_{\gamma0 } } \left[\frac{3 \left ( 100~{\rm km~s}^{-1}~{\rm mpc}^{-1}\right)^2 } { 8 \pi g_{\rm n } m_{\rm h } } \right ] \frac{m_{\rm h}}{m_{{\rm b}0 } } \nonumber\\ & = & 2.7378 \times 10^{-8 } \left(\frac{g_{\rm n}}{6.6738\times 10^{-11}~{\rm m}^3~{\rm kg}^{-1}~{\rm s}^{-2}}\right)^{-1 } \left ( \frac{t_{\gamma 0}}{2.7255~{\rm k}}\right)^{-3 } \left [ 1 + 0.007131 \left ( y_{\rm p } -0.25\right ) \right ] .
\label{eta_omegab_relation}\end{aligned}\ ] ] we adopt the cmb temperature @xcite @xmath70 the latest gravitational constant @xcite is given by @xmath71 the square - root of sum of squares of uncertainties from @xmath62 and @xmath72 is taken , and the precise relation between @xmath4 and @xmath73 is derived , for @xmath74 , as @xmath75 it is assumed that energetic electrons and positrons are generated at the decay of a long - lived massive particle ( sterile neutrino ) @xmath1 .
the sterile neutrino @xmath1 has a mass @xmath76 and a mean lifetime @xmath77 . through the inverse compton scattering of cosmic background radiation ( cbr ) , the energetic electrons and positrons produce energetic primary photons .
the energies of the photons are related to the energy of @xmath3 and temperature of the universe .
the present model has three parameters regarding effects of nonthermal nucleosynthesis : ( 1 ) @xmath78 is the number ratio of the decaying particle @xmath1 and the background photon evaluated at a time between the cosmological electron positron pair annihilation and the @xmath1 decay , ( 2 ) the @xmath1 mass @xmath76 , and ( 3 ) the lifetime @xmath77 .
the total energy of @xmath3 emitted at one @xmath1 decay event @xmath79 is derived as a function of @xmath76 .
this quantity is equivalent to the energy injected in the form of electromagnetic cascade showers .
we adopt the method of ref .
@xcite to calculate the nonthermal nucleosynthesis , where thermonuclear reactions are simultaneously solved .
we utilized updated reaction rates of @xmath34he photodisintegration ( eqs .
( 2 ) and ( 3 ) of ref .
@xcite ) based on cross section data from measurements with laser - compton photons @xcite . in this work , we found errors in cross sections of reactions @xmath6be(@xmath11 , @xmath12)@xmath13he and @xmath6li(@xmath11 , @xmath12)@xmath13h @xcite adopted in previous studies on the bbn model with the long - lived decaying particle .
the errors are corrected in this calculation as explained in sec .
[ sec6 ] .
the injection spectrum of nonthermal photon is given by @xmath80 where @xmath81 is the photon temperature of the universe , @xmath82 is the electron mass , @xmath83 is the differential decay rate as a function of energy of @xmath3 and @xmath76 [ cf .
( [ eq_a15 ] ) ] , @xmath84 is the energy spectrum of primary photon ( @xmath85 ) produced at inverse compton scatterings of @xmath3 with energy @xmath86 at temperature @xmath81 , @xmath87 is the energy spectrum of nonthermal photon with energy @xmath88 which is produced in the electromagnetic cascade showers triggered by primary photon with energy @xmath89 at @xmath81 , and @xmath90 is the maximum value of @xmath89 [ eq .
( [ e_g0_max ] ) ] . throughout this paper ,
we concentrate on the dirac neutrino case .
we denote the mass of the heaviest state of neutrinos as @xmath76 and the active - sterile mixing angle as @xmath91 . the energy spectrum of @xmath3 , i.e. , @xmath92 is calculated ( appendix [ app1 ] ) , and input in eq .
( [ eq_inj ] ) . in this paper
, we fix parameters of the sterile neutrino as one possible simple example case as follows .
the sterile neutrino couples to charged and neutral currents of the electron flavor only .
the strengths of the @xmath1 coupling to the currents are given by @xmath91 ( for electron ) and @xmath93 ( for muon and tauon ) .
we note that the investigation in this paper can be easily extended to the majorana case .
the only difference between the two cases is that decay rates of a majorana sterile neutrino are twice as large as those of a dirac neutrino when those neutrinos have the same parameter values of @xmath76 and @xmath91 .
the relevant lagrangian of neutrino sector including one majorana sterile neutrino is given by @xmath94 where @xmath95 and @xmath96 are lepton and higgs doublets , respectively , @xmath97 ( for @xmath98 ) is the yukawa coupling constant , and @xmath76 is the majorana mass . after the breaking of ew symmetry , the sterile neutrino is mixed with active neutrinos .
the degree of mixing is characterized by the active - sterile mixing denoted as @xmath99 , where @xmath100 denotes the vacuum expectation value of the higgs field .
the sterile neutrino can decay through this mixing into @xmath101 or @xmath102 , where @xmath103 is a charged lepton species .
inversely , if sterile neutrino is relatively light , it can be produced by the decay of light mesons .
for instance when @xmath104 ( pion mass ) is satisfied , the pion decay can produce the sterile neutrino . in this case
, we can find a signal of sterile neutrino as a new peak in the energy spectrum of active neutrinos .
we assume that energetic electrons and positrons , @xmath3 s , are injected at the @xmath1 decay .
energetic photons , @xmath11 s , are then produced via the inverse - compton scattering between the @xmath3 s and cbr ( @xmath105 ) , where the subscript `` bg '' means background .
the inverse - compton process plays an important role in developments of electromagnetic cascade shower .
it distributes the energy of @xmath3 generated at the decay to multiple particles , i.e. , @xmath11 and @xmath3 in the thermal bath .
we assume that the @xmath3 in the initial state with energy @xmath86 reacts with cbr with energy @xmath106 , and a scattered photon in the final state has energy @xmath89 .
number of collisions for an @xmath3 particle per unit time @xmath107 and unit energy interval of photon in the final state @xmath108 is then approximately given @xcite by @xmath109 where @xmath110 is the classical radius of electron with the fine structure constant @xmath12 , and the function @xmath111 is defined by @xmath112 where parameters are introduced as @xmath113 the energy spectrum of cbr in the early universe is almost completely described by a planck distribution , @xmath114 generally , at inverse compton scatterings of low energy cbrs by energetic @xmath3 in the early universe , small fractions of energies of @xmath3 are transferred to those of cbrs at respective scatterings , i.e. , @xmath120 .
the energy spectrum of the primary photon produced at the inverse compton scattering is , therefore , approximately proportional to the differential scattering rate as a function of @xmath89 .
the spectrum of the primary photon is then given by @xmath121 } , \label{eq_p_ic}\ ] ] where li@xmath122 is the dilogarithm .
the dilogarithm is calculated in our code using the published algorithm 490 @xcite .
the energetic primary photons interact with the cbr , and electromagnetic cascade showers composed of energetic photons and @xmath3 develop ( e.g. , @xcite ) .
one energetic photon can produce multiple particles of lower energies by continuous reactions of pair production at a collision with cbr ( @xmath123 ) and the inverse compton scattering of @xmath3 at a collision with cbr ( @xmath124 ) .
the nonthermal photons then obtain a quasi - static equilibrium spectrum @xcite .
when the energy of the primary photon @xmath89 is much larger than the threshold energy for photodisintegration of light nuclides , i.e. , @xmath125 mev , the steady state energy spectrum of the nonthermal photons is approximately given ( e.g. , @xcite ) by @xmath126 where @xmath127 and @xmath128 @xcite are the energy corresponding to a break in the power law , and a cutoff energy , respectively , @xmath129\}$ ] is the normalization constant which conserves the energy of the primary photon .
if nonthermal photons have energies larger than @xmath130 , they are quickly destroyed via the electron - positron pair production . the maximum energy of the nonthermal photon @xmath89 , however , can be of the order of @xmath131(1 mev ) depending on the mass @xmath76 and the temperature @xmath81 [ eq .
( [ e_g0_max ] ) ] . in this case , the generalized photon spectrum is given @xcite as follows : ( 1 ) for @xmath132 , the spectrum is given by @xmath133 where @xmath134 .
+ ( 2 ) for @xmath135 , the spectrum is given by @xmath136 where @xmath137\}$ ] .
+ ( 3 ) for @xmath138 , the spectrum @xmath139 is given by eq . ( [ eq_spe ] ) .
rates of electromagnetic interactions are faster than the cosmic expansion rate .
the injection spectrum @xmath140 is then quickly modified to a new quasi - static equilibrium spectrum given by @xmath141 where @xmath142 is the number density of the decaying particles @xmath1 at a redshift @xmath143 , and @xmath144 is a parameter describing the amount of electromagnetic energy injection , with @xmath145 and @xmath146 the comoving number densities of @xmath1 and cbr , respectively , estimated at a time between the cosmological @xmath3 annihilation and the @xmath1 decay . here ,
the cosmic time is described as @xmath147 taking the inverse function of the temperature evolution @xmath148 .
the quantity @xmath149 is the energy degradation rate of the nonthermal photons through three relatively slow processes ; compton scattering ( @xmath150 ) , bethe - heitler ordinary pair creation in nuclei ( @xmath151 ) , and double photon scattering ( @xmath152 ) , where @xmath153 is a nucleon in nucleus .
since the energy degradation rate depends not only on the photon temperature but also on the baryon and electron densities @xcite , it is a function of the baryon - to - photon ratio @xmath4 .
we fix the @xmath4 value after the @xmath1 decay to the observed value from the planck cbr measurement .
the entropy production associated with the @xmath1 decay changes the @xmath4 value as a function of time .
therefore , the time evolution of the baryon - to - photon ratio depends on the parameters , i.e. , @xmath154 .
we use this steady state approximation for the nonthermal photon spectrum .
if the injection of nonthermal photons occurs at a cosmic time @xmath155 s , the nonthermal photons can disintegrate background nuclei , and nuclear abundances can be changed @xcite .
the equation describing a time evolution of nuclear abundance by primary reactions , i.e. , photodisintegration , is given by @xmath156_a y_t - \sum_p \left[a\gamma\right]_p y_a , \label{dydt2}\ ] ] where @xmath157 is the mole fraction of a nuclear species @xmath18 with @xmath17 and @xmath158 number densities of nuclide @xmath18 and total baryon , respectively , .
equations ( [ dydt2])-([tt ] ) are exact for a general case including the energy injection .
equations for @xmath159 in ref . @xcite and @xmath160 in ref .
@xcite are , on the other hand , not exact when the effect of the electromagnetic energy injection by the exotic particle decay is significant . ] and we define the reaction rate @xmath161_a = \frac{\eta_{{\nu_{\rm h}}}(t ; m_{{\nu_{\rm h } } } , \tau_{{\nu_{\rm h } } } , \zeta_{{{\nu_{\rm h}}}\rightarrow e } ) e_{{{\nu_{\rm h}}}\rightarrow e } } { \tau_{{\nu_{\rm h } } } } ~g_{[p\gamma]_a}^1(t ; m_{{\nu_{\rm h } } } , \tau_{{\nu_{\rm h } } } , \zeta_{{{\nu_{\rm h}}}\rightarrow e } ) , \label{nonthermal_rate1}\ ] ] where @xmath162 is the @xmath1-to - photon number ratio with @xmath163 the number density of cbr , and the nuclear transfer function and the normalized spectrum of nonthermal photon are defined as @xmath164_a}^1(t ; m_{{\nu_{\rm h } } } , \tau_{{\nu_{\rm h } } } , \zeta_{{{\nu_{\rm h}}}\rightarrow e } ) = \int_{0}^\infty de_\gamma t_\gamma^{\rm qse}(e_\gamma ; t , m_{{\nu_{\rm h } } } , \tau_{{\nu_{\rm h } } } , \zeta_{{{\nu_{\rm h}}}\rightarrow e})\ , \sigma_{\gamma + p \rightarrow a } ( e_\gamma ) , \label{agamma}\ ] ] and @xmath165(e_\gamma ; t , m_{{\nu_{\rm h } } } , \tau_{{\nu_{\rm h } } } , \zeta_{{{\nu_{\rm h}}}\rightarrow e } ) } .
\label{sgamma}\end{aligned}\ ] ] the first and second term on the right hand side ( rhs ) in eq .
( [ dydt2 ] ) correspond to the production ( @xmath166 ) and destruction ( @xmath167 ) , respectively , for nuclide @xmath40 .
the cross section of the process @xmath168 is denoted by @xmath169 .
in addition , we use @xmath170 as the number of identical nuclear species in a production process ; @xmath171 when particles @xmath40 and @xmath172 are identical and @xmath173 when they are not .
for example , in the process @xmath34he(@xmath11 ,
@xmath174)d , @xmath175 since two deuterons are produced at one reaction .
nonthermal nuclei produced in the primary reaction can in general experience secondary nonthermal nuclear reactions .
we , however , focus on an injection of photons with relatively small energies generated by a light sterile neutrino of @xmath176 mev . in this case , @xmath34he photodisintegration is impossible , and most of important secondary reactions do not occur . in this calculation , only the secondary reactions @xmath6li(@xmath11 , @xmath177)@xmath178li(@xmath179 , @xmath12)@xmath13he and @xmath6be(@xmath11 , @xmath179)@xmath178li(@xmath179 , @xmath12)@xmath13he are operative .
the equation describing the secondary production and destruction is given by @xmath180_s -({\rm destruction~term } ) , \label{dysdt}\ ] ] where the reaction rate for a secondary reaction @xmath181 with any combination of particles @xmath182 , @xmath40 , and @xmath183 is given by @xmath184_s = \frac{\eta_{{\nu_{\rm h}}}(t ; m_{{\nu_{\rm h } } } , \tau_{{\nu_{\rm h } } } , \zeta_{{{\nu_{\rm h}}}\rightarrow e } ) e_{{{\nu_{\rm h}}}\rightarrow e } } { \tau_{{\nu_{\rm h } } } } ~g_{[p(a)p']_s}^2(t ; m_{{\nu_{\rm h } } } , \tau_{{\nu_{\rm h } } } , \zeta_{{{\nu_{\rm h}}}\rightarrow e } ) , \label{nonthermal_rate2}\ ] ] @xmath185_s}^2(t ; m_{{\nu_{\rm h } } } , \tau_{{\nu_{\rm h } } } , \zeta_{{{\nu_{\rm h}}}\rightarrow e } ) & = & \int_{0}^\infty de_a \frac{\sigma_{a+p'\rightarrow s}(e_a)\beta_a(e_a ) } { \left [ b_a / n_b \right](e_a ; t , m_{{\nu_{\rm h } } } , \tau_{{\nu_{\rm h } } } , \zeta_{{{\nu_{\rm h}}}\rightarrow e } ) } \nonumber \\ & & \times \int_{{\mathcal e}_a^{-1}(e_a)}^\infty de_\gamma t_\gamma^{\rm qse } ( e_\gamma ; t , m_{{\nu_{\rm h } } } , \tau_{{\nu_{\rm h } } } , \zeta_{{{\nu_{\rm h}}}\rightarrow e } ) \sigma_{\gamma + p\rightarrow a } ( e_\gamma ) \nonumber \\ & & \times \exp{\left [ -\int_{e_a}^{{\mathcal e}_a(e_\gamma ) } de_a ' \frac{\gamma_a(e_a ' ) } { b_a(e_a ' ; t , m_{{\nu_{\rm h } } } , \tau_{{\nu_{\rm h } } } , \zeta_{{{\nu_{\rm h}}}\rightarrow e } ) } \right ] } , \label{tt'}\end{aligned}\ ] ] where @xmath186 is the velocity of the nuclide produced in a primary reaction , i.e. , primary nuclide , @xmath40 , @xmath187 is the energy loss rate of the primary nuclide mainly from coulomb scattering of electrons , and @xmath188 is the destruction rate of the primary nuclide .
stable nuclides have @xmath189 , while unstable nuclides have nonzero values given by the @xmath190-decay rates .
the quantity @xmath191 is the energy of the nuclide @xmath40 produced at the reaction @xmath192 of nonthermal photon with energy @xmath88 , and @xmath193 is the energy of the nonthermal photon which produces the primary nuclide @xmath40 with energy @xmath194 . the transfer functions @xmath195_a}^1(t ; m_{{\nu_{\rm h } } } ,
\tau_{{\nu_{\rm h } } } , \zeta_{{{\nu_{\rm h}}}\rightarrow e})$ ] and @xmath196_s}^2(t ; m_{{\nu_{\rm h } } } , \tau_{{\nu_{\rm h } } } , \zeta_{{{\nu_{\rm h}}}\rightarrow e})$ ] should be derived as a function of @xmath81 for a fixed parameter set of ( @xmath76 , @xmath197 , @xmath198 ) .
time evolutions of the temperature @xmath199 , the baryon - to - photon ratio @xmath200 , and the @xmath1-to - photon ratio @xmath201 are different for different parameter sets . in eq .
( [ agamma ] ) , the parameter dependence of the transfer function comes from the steady state nonthermal photon spectrum @xmath202 and the energy loss rate per background photon @xmath203(e_\gamma ; t , m_{{\nu_{\rm h } } } , \tau_{{\nu_{\rm h } } } , \zeta_{{{\nu_{\rm h}}}\rightarrow e})$ ] . in eq .
( [ tt ] ) , @xmath204 is independent of the baryon to photon ratio @xmath4 at the time since the energy loss rate of energetic nuclide , @xmath205 , is proportional to the baryon density @xcite .
however , it depends on the @xmath4 value at bbn through the @xmath34he abundance ( or the electron abundance ) .
then , the quantity @xmath204 , the photon spectrum @xmath206 , and the energy loss rate @xmath205 in the exponential term depend on the parameter set ( @xmath207 ) . in this study
, we calculate the transfer functions neglecting the entropy production by the sterile neutrino decay .
then , the @xmath4 value is constant , and set to be consistent with the central value determined by planck .
also the @xmath1-to - photon ratio is exactly described as @xmath208 [ cf .
( [ n_nu_h ] ) ] . in this case
, the transfer functions depend on the temperature @xmath81 and the mass @xmath76 only .
the reaction rates [ eqs . ( [ nonthermal_rate1 ] ) and ( [ nonthermal_rate2 ] ) ]
then reduce to @xmath209_a & = & \frac{\zeta_{{{\nu_{\rm h}}}\rightarrow e}}{\tau_{{\nu_{\rm h } } } } \exp(-t/\tau_{{\nu_{\rm h}}})~g_{[p\gamma]_a}^1(t ; m_{{\nu_{\rm h } } } ) , \label{nonthermal_rate1b } \\
\left[p(a)p'\right]_s & = & \frac{\zeta_{{{\nu_{\rm h}}}\rightarrow e}}{\tau_{{\nu_{\rm h } } } } \exp(-t/\tau_{{\nu_{\rm h}}})~g_{[p(a)p']_s}^2(t ; m_{{\nu_{\rm h } } } ) .
\label{nonthermal_rate2b}\end{aligned}\ ] ] we calculate nonthermal nucleosynthesis triggered by the nonthermal photons taking @xmath77 , @xmath198 , and @xmath76 as parameters . by using these simplified transfer functions ,
parameter search can be performed with same transfer functions for respective @xmath76 values in realistic computation time .
we can safely use the simplified transfer functions without missing a parameter region in which the primordial @xmath6li abundance is reduced , for the following reason . in this model
, the entropy production by the @xmath1 decay always reduces the baryon - to - photon ratio as a function of time ( sec .
[ sec4 ] ) .
the ratio in the bbn epoch is , therefore , larger than that in the cosmological recombination epoch measured by planck .
then , one can place a lower limit on the bbn @xmath4 value taking the planck value . on the other hand ,
deuterium is only destroyed by photodisintegration , and is never produced in the nonthermal nucleosynthesis by the @xmath1 decay .
the abundance of deuterium produced at the bbn , therefore , can not be significantly smaller than the observational constraint on the primordial abundance .
this requirement gives an upper limit on the bbn @xmath4 value since primordial d abundance decreases as a function of @xmath4 .
these limits are satisfied in a very narrow region of the bbn @xmath4 value , only @xmath45(1 ) % wide , which is around the planck value ( see fig . 1 of ref
. @xcite or fig .
1 of ref , @xcite ) . then , maximum allowed changes of the @xmath4 value between the bbn and the recombination epochs is @xmath45(1 ) % . also the @xmath210 value in the case with an entropy production can change from that in the case without it by only @xmath45(1 ) % . the neglect of the entropy production effect in eqs .
( [ nonthermal_rate1b ] ) and ( [ nonthermal_rate2b ] ) , therefore , does not introduce a large error in final nuclear abundances . in the following calculation ,
the final @xmath4 value is fixed to the planck value .
when the entropy production changes the @xmath4 and @xmath210 values by more than @xmath35 % , present results of the nonthermal nucleosynthesis calculation are not precise .
however , such a large entropy production is accompanied by a large bbn @xmath4 value , and is therefore excluded by underproduction of deuterium definitely .
we correct significant errors in cross sections of reactions @xmath6be(@xmath11 , @xmath12)@xmath13he and @xmath6li(@xmath11 , @xmath12)@xmath13h @xcite adopted in previous studies on the bbn model including a decaying particle ( e.g. , @xcite ) .
the detailed balance relation between cross sections of a forward reaction @xmath40(@xmath43 , @xmath11)@xmath172 and its inverse reaction @xmath172(@xmath11 , @xmath43)@xmath40 is described @xcite as @xmath211 where @xmath212 and @xmath213 are the forward and inverse reaction cross sections , respectively , @xmath214 is the statistical degrees of freedom ( dof ) with spin @xmath215 of species @xmath18 , @xmath216 is the kronecker delta for avoiding the double counting of identical particles , @xmath217 and @xmath218 are the reduced mass and the center of mass ( cm ) energy , respectively , of the @xmath219 system , and @xmath220 is the radiation energy with @xmath221 the reaction @xmath221-value : @xmath222 mev [ for @xmath13he(@xmath12 , @xmath11)@xmath6be ] and @xmath223 mev [ for @xmath13h(@xmath12 , @xmath11)@xmath6li ] , respectively . the forward reaction rate is described using the astrophysical @xmath224-factor as @xmath225{\mathstrut \frac{e_{\rm g}}{e } } \right ) , \label{eq_b2}\ ] ] where @xmath226 is the gamow energy with @xmath227 the proton number of species @xmath18 and @xmath228 the fine structure constant . inserting this equation in eq .
( [ eq_b1 ] ) , we obtain a relation between the cross section of the inverse ( photodisintegration ) reaction and the @xmath224-factor of the forward ( radiative capture ) reaction : @xmath229{\mathstrut \frac{e_{\rm g}}{e } } \right ) .
\label{eq_b3}\ ] ] firstly , @xmath224-factors of the two reactions are taken from ref .
@xcite . when we take fitted functions [ their eqs .
( 6 ) and ( 7 ) ] with theoretical values of @xmath230 kev b for @xmath6be(@xmath11 , @xmath12)@xmath13he and @xmath231 kev b for @xmath6li(@xmath11 , @xmath12)@xmath13h , the photodisintegration cross sections are given by @xmath232 where @xmath233 and @xmath234 are defined , and @xmath235 it is found that both cross sections are smaller than the published values @xcite by a factor of about three although the values were based on the same reference @xcite .
it is expected that this error originates from wrong treatment of statistical dof .
another error stems from the use of the fitted functions @xcite which can be reasonably applied only to the low energy region of @xmath236 mev .
the functions have been derived from fitting to measured data in a relatively low energy region .
these functions then provide erroneous values in a large @xmath88 region .
especially , the function for @xmath6be(@xmath11 , @xmath12)@xmath13he outputs large negative value at @xmath237 mev . since this error affects the final nuclear abundances calculated in the model with a decaying particle , it must be fixed .
for example , we may assume constant @xmath224-factors : @xmath238 kev b at @xmath239 mev for @xmath6be(@xmath11 , @xmath12)@xmath13he and @xmath240 kev b at @xmath241 mev for @xmath6li(@xmath11 , @xmath12)@xmath13h ( cf . ref .
the cross sections are then given by @xmath242 @xmath243 recently , three independent groups have measured the cross section of the reaction @xmath13he(@xmath12 , @xmath11)@xmath6be at high energies ( @xmath244 mev ) with relatively small errors @xcite .
there is only one earlier publication for the cross section measured in this energy region , and the measured data included large errors @xcite .
the new measurements indicated a possible increase of the cross section at @xmath244 mev which was not seen in the earlier work .
theoretically , this increase can be understood as a contribution of the electric dipole capture from the scattering @xmath174-wave which gradually becomes more important at higher energies @xcite .
when we consider these experimental results , it may be better to take into account the @xmath174-wave behavior in an estimation of the cross section at high energies rather than to assume a constant @xmath224-factor .
we note , however , that a precise estimation needs more measurements at high energy of @xmath245 mev .
the only experimental data in the high energy region are from measurements for the @xmath13he incident energy @xmath246 mev at the angle 90@xmath247 @xcite .
the measured data typically give only upper limits on the differential cross section , @xmath248b / sr , at the corresponding cm energy @xmath249 mev .
an analytical function has been fitted to cross sections measured between 2004 and 2007 in the energy range of @xmath250 mev @xcite .
although this fitting did not take into account experimental data at higher energies , dominant contributions of @xmath251- and @xmath174-waves are included in the analytical function .
we then adopt this cross section function with six parameters in this paper .
the photodisintegration cross section is given by @xmath252 , \label{eq_b8}\end{aligned}\ ] ] where @xmath253 and 1 indicate capture cross sections to the ground and the first excited states of @xmath6be , respectively , and @xmath254 and @xmath255 mev are the @xmath221-values for the two final states .
the six fitted parameters are @xmath256 , @xmath257 , @xmath258 , @xmath259 , @xmath260 , and @xmath261 .
figure [ fig_a1 ] shows cross sections of reactions @xmath6be(@xmath11 , @xmath12)@xmath13he and @xmath6li(@xmath11 , @xmath12)@xmath13h as a function of the photon energy .
the cross sections are derived by applying the detailed balance relation to the forward radiative capture cross sections .
solid and dashed lines correspond to polynomial fits to theoretical calculations in low energy regions @xcite and constant @xmath224-factors in high energy regions @xcite , respectively .
the dot - dashed line corresponds to fit to experimental data on @xmath13he(@xmath12 , @xmath11)@xmath6be
@xcite . for comparison ,
dotted lines show fitted functions of ref .
we adopt the dot - dashed line [ eq .
( [ eq_b8 ] ) ] for @xmath6be(@xmath11 , @xmath12)@xmath13he , and the solid and the dashed lines [ eqs .
( [ eq_b5 ] ) and ( [ eq_b7 ] ) ] for @xmath6li(@xmath11 , @xmath12)@xmath13h in the following calculations .
be(@xmath11 , @xmath12)@xmath13he and @xmath6li(@xmath11 , @xmath12)@xmath13h as a function of the photon energy .
they are estimated from the detailed balance relation with the forward radiative capture cross sections .
solid lines correspond to polynomial fits to theoretical calculations in the low energy regions @xcite .
dashed lines correspond to constant @xmath224-factors in the high energy regions @xcite .
the dot - dashed line is from a fit to experimental data on @xmath13he(@xmath12 , @xmath11)@xmath6be @xcite .
dotted lines show fitted functions of ref .
[ fig_a1],width=302 ]
the decay of sterile neutrino has two effects on resulting effective neutrino number in the universe , i.e. , @xmath5 .
firstly , since the decay generates energetic neutrinos , the total energy density of neutrino increases with respect to the case of no neutrino injection @xcite . secondly , since the decay also generates energetic electron and positron , the energy density of background photon increases @xcite .
the ratio between energy densities of neutrino and photon is reduced , and as a result , effective neutrino number is reduced compared to the case of no photon injection .
the baryon - to - photon ratio is simultaneously reduced by this photon heating effect . in this study , we assume that the lifetime of the sterile neutrino is much larger than the time scale of the decoupling of the active neutrino in the early universe , i.e. , @xmath262 s. the energetic neutrino generated at the decay can not interact effectively with background particles mainly constituted of weakly interacting neutrinos and weakly - noninteracting photon .
the energetic neutrino is then never thermalized , and propagates in the universe without collisions .
the total energy density in the universe is given @xcite by @xmath263 where @xmath264 is the energy density of particle species @xmath265 ( photon ) , @xmath266 ( electron ) , @xmath267 ( positron ) , @xmath268 ( thermal neutrino ) , @xmath42 ( baryon ) , @xmath1 ( sterile neutrino ) , and @xmath269 ( nonthermal neutrino produced at the @xmath1 decay )
. the last two terms are not present unless the decaying sterile neutrino exists .
the sterile neutrino energy density is given @xcite by @xmath270 where @xmath17 is the energy density of particle @xmath18 , and @xmath271 is the energy density of @xmath18 at time @xmath272 ( @xmath273 mev ) . the initial number density of @xmath1 is related as @xmath274 the factor ( 11/4 ) originates from the entropy transfer from @xmath3 to photons at the cosmological @xmath3 annihilation .
when the lifetime of the sterile neutrino is equal to or longer than the bbn time scale , the cosmic expansion rate can be affected by the energy density of the sterile neutrino .
since such a change in expansion rate changes resultant elemental abundances , it is constrained from observed light element abundances . for example , the energy density of exotic relativistic species in the bbn epoch has been constrained @xcite .
the sterile neutrino with a mass of @xmath275 mev is nonrelativistic during bbn .
the energy density of a nonrelativistic particle redshift as @xmath276 , while that of a relativistic particle redshifts as @xmath277 .
therefore , the former increases relative to the latter as the universe expands . the effect of the nonrelativistic sterile neutrino is thus different from that of the exotic relativistic particle .
the change in the expansion rate during bbn is rigorously taken into account by using eq .
( [ eq_rho_tot ] ) .
the nonthermal neutrino energy density is solved by time integration @xcite of @xmath278 where @xmath279 is the average total energy of active neutrinos and antineutrinos emitted per one @xmath1 decay event , and the cosmic expansion rate is given by @xmath280 where @xmath41 is the scale factor of the universe
. the integral form of eq .
( [ drho_nudt ] ) is given by @xmath281 where @xmath282 is the initial energy density of @xmath1 , and @xmath283 is the scale factor at time @xmath272 .
we used eq .
( [ eq_n_nuh_i ] ) at the second equality .
the ratio of the average total energy emitted in the form of electron and positron at the @xmath1 decay to the mass @xmath76 is given by @xmath284 where @xmath285 is the ratio of average total energies injected in the forms of @xmath286 ( including all flavors and antineutrinos ) and @xmath3 at the @xmath1 decay .
in the present model , the sterile neutrino has only two decay modes , i.e. , @xmath287 and @xmath288 .
the ratio of the decay rates for the two modes is defined with eqs .
( [ eq_a9 ] ) and ( [ eq_a21 ] ) as @xmath289 through the two decay modes , energetic neutrinos , electrons , and positrons are generated .
the ratio of average total energies of @xmath286 and @xmath3 emitted through the two decay modes is given by @xmath290 } , \label{eq_ratio_nue}\ ] ] where @xmath291 is the ratio of the average energy of species @xmath18 to the sterile neutrino mass in the decay mode of @xmath287 , and the equation , @xmath292 is satisfied .
there is a trivial relation between parameters : @xmath293 where @xmath294 is defined .
figure [ fig8 ] shows the ratio of the decay rates @xmath295 as a function of @xmath296 .
when the sterile neutrino mass is much larger than the electron mass , i.e. , @xmath297 , the rate for the decay into @xmath298 is comparable to that into three neutrinos .
the ratio monotonically decreases with increasing @xmath299 .
when the sterile neutrino mass is nearly one half of the electron mass , i.e. , @xmath300 , this ratio becomes very small and the decay into three neutrinos dominates . as a function of @xmath296 .
[ fig8],width=302 ] figure [ fig9 ] shows the ratio of average total energies of neutrinos and @xmath3 s injected at the sterile neutrino decay @xmath285 as a function of @xmath299 . in the large limit of the sterile neutrino mass , @xmath301 , the ratio approaches to @xmath302 .
the ratio monotonically increases with increasing @xmath299 .
the ratio diverges in the small mass limit , @xmath303 .
s injected at the sterile neutrino decay @xmath285 [ eq .
( [ eq_ratio_nue ] ) ] as a function of @xmath296 .
[ fig9],width=302 ] equation of the energy conservation is given ( eq .
( d.26 ) of ref .
@xcite ) by @xmath304 where @xmath179 is the total pressure of the universe , and the third and fourth terms in rhs correspond to the energy changes by nucleosynthesis and the electromagnetic energy injection at the sterile neutrino decay , respectively .
the term of the electromagnetic energy injection is given by @xmath305 the equation of the energy conservation is then transformed to @xmath306 where @xmath307 is defined , @xmath308 , and @xmath309 and @xmath310 are the total energy and pressure , respectively , of electron and positron .
this is a varied form of the temperature evolution as a function of time , @xmath311 ( cf .
( 22 ) of ref .
@xcite ) , and is used in kawano s bbn code .
the last term in the brackets in the denominator of rhs is absent unless the entropy production is induced by the @xmath1 decay ( cf .
( d.28 ) of ref .
@xcite ) .
the effective neutrino number is defined by @xmath312 where @xmath313 is the total neutrino energy density .
the thermal neutrinos interact with thermal bath until the temperature decreases to @xmath314 mev .
then , they decouple from background photons at @xmath315 mev , and have temperature @xmath316 which can be different from the photon temperature @xmath81 .
the energy density of the thermal neutrinos is given by @xmath317 where @xmath318 is the effective neutrino number at the initial time @xmath319 mev before the cosmological @xmath3 annihilation .
the nonthermal neutrinos are assumed to originate only from the @xmath1 decay .
since we assume that the decay occurs much later than the bbn epoch , nonthermal neutrinos from the decay can not be thermalized typically .
the nonthermal neutrinos , therefore , need to be treated separately from the thermal ones .
the photon number density is given by @xmath320 using eqs .
( [ eq_rho_nu_nt ] ) , ( [ eq_neff ] ) , ( [ eq_rho_nu_th ] ) , and ( [ eq_rho_g ] ) , the effective neutrino number is transformed to @xmath321 we use @xmath322 and @xmath323constant . the effective neutrino number can then be described as @xmath324 .
\label{eq_neff_final}\ ] ] the factor @xmath325 in rhs results from the entropy production by the injection of nonthermal electron and positron at the @xmath1 decay .
the number and energy densities of thermal neutrinos for a fixed photon temperature @xmath81 are decreased compared to the case of no entropy production .
the second term in the square brackets corresponds to an increase of the neutrino energy density contributed by the nonthermal active neutrino injection .
the time derivative of the effective neutrino number can be derived from this equation as @xmath326 then , we solve the time evolution of the effective neutrino number simultaneously with those of @xmath327 [ eq . ( [ hubble ] ) ] , @xmath199 [ eq . ( [ drdt9 ] ) ] , and @xmath328
. an injection of energetic @xmath3 at the @xmath1 decay produces nonthermal photons via electromagnetic cascade showers , and enhances the comoving photon entropy in the universe .
since the baryon - to - photon ratio @xmath4 is inversely proportional to the comoving photon entropy , the ratio is reduced as a function of time during the nonthermal photon injection @xcite . in this case , the @xmath4 value in the bbn epoch is larger than that in the epoch of the cbr last scattering .
we adopt the baryon - to - photon ratio inferred from planck measurement of cbr as the value after the @xmath1 decay .
the initial @xmath4 value is then determined as a function of @xmath77 and @xmath198 such that it results in the final @xmath4 value consistent with the planck data .
here we define the following variables : the time @xmath329 is between the cosmological @xmath3 annihilation epoch and the @xmath1 decay , while the time @xmath330 is after the decay .
the quantities @xmath331 and @xmath332 are the comoving photon entropies at times @xmath329 and @xmath330 , respectively , while @xmath333 and @xmath334 are baryon - to - photon ratios at times @xmath329 and @xmath330 , respectively .
the comoving entropy density is given @xcite by @xmath335 where @xmath336 is the relativistic dof in terms of entropy .
since we consider the @xmath1 decay after the neutrino decoupling at @xmath314 mev , the photon is the only relativistic component in the thermal bath .
therefore , this dof is the same as the statistical dof of photon , i.e. , @xmath337 .
the scale factor of the universe is inversely proportional to the neutrino temperature .
the ratio of the comoving photon entropies @xmath338 is , therefore , given by @xmath339 where @xmath340 is the neutrino temperature before the @xmath1 decay , and we used the relation @xmath341 .
the planck measurement has obtained the @xmath342 value from combined data of planck+wp+highl+bao : @xmath343 ( 68% c.l . ) @xcite .
therefore , the error in the baryon - to - photon ratio [ eq .
( [ eta_omegab_relation ] ) ] or the comoving entropy at the cbr last scattering is lower than 2.2% ( @xmath344 ) . when the comoving photon entropy changes by a small fraction , i.e. , @xmath345 % in the epoch of the @xmath1 decay , the ratio @xmath346 is approximately given @xcite by @xmath347,\ ] ] where @xmath348 and @xmath349 are the relativistic dofs in terms of energy and entropy , respectively , after the bbn epoch , @xmath350 and @xmath351 are number densities of the decaying particle and photon , respectively , evaluated at the same time @xmath329 , and @xmath352 gev is the planck mass .
the @xmath353 value is the same as statistical dof of photon . and
photon itself during the cosmic epoch after bbn .
this is because light active neutrinos had been decoupled from thermal bath before the onset of bbn at @xmath314 mev .
we note that previous studies ( e.g. , @xcite ) are erroneous since the @xmath354 factor included the contribution from neutrinos . the right numerical value in the following eq .
( [ eq_entropy ] ) is then higher than the previous suggestion by a factor of 3.91/2=1.96 . ] for a small fractional change of entropy , this value is given @xcite by @xmath355 the ratio of the baryon - to - photon ratios @xmath356 can be estimated using this formula .
the energy density of nonthermal active neutrino is generated by the @xmath1 decay , and an approximate formula can be derived as in the case of a nonthermal photon injection in the late universe @xcite .
it is assumed that the sterile neutrino does not contribute to the total energy density so much that the cosmic expansion rate is not affected much .
the standard radiation dominated universe then holds .
we define the second term in the square brackets of eq .
( [ eq_neff_final ] ) as @xmath357 .
this term then approximately becomes @xmath358 where the effective time is defined as @xmath359^{1/\beta } \tau_{{\nu_{\rm h}}}$ ] for the universe with the time - temperature relation of @xmath360 with @xmath361 the gamma function of argument @xmath362.^{1/\beta } \tau_{{\nu_{\rm h}}}$ ] in ref .
] for the radiation dominated universe considered here , @xmath363 and @xmath364 are satisfied .
if the nonthermal neutrino injection is not accompanied by the nonthermal photon injection , the photon heating never occurs and the effective neutrino number after the decay is @xmath365 . on the other hand , when nonthermal electrons and positrons are injected at the @xmath1 decay
, the entropy of the universe is increased .
the number and energy densities of neutrino for a fixed photon temperature @xmath81 are then decreased compared to the case of no entropy production .
the effective number after the decay is given @xcite by @xmath366 an approximate solution of @xmath5 is then derived with eqs .
( [ eq_entropy ] ) , ( [ neff_nt_i ] ) , and ( [ eq_dn ] ) .
we constrain the model of the decaying sterile neutrino by comparing calculated results and observational constraints on elemental abundances , the effective neutrino number , and the cmb energy spectrum . in this model ,
@xmath6be is disintegrated by nonthermal photons originating from the @xmath1 decay . since a primordial abundance of @xmath6li
is mainly contributed from that of @xmath6be produced during the bbn epoch , the destruction of @xmath6be reduces the final @xmath6li abundance .
however , the energies of nonthermal photons should be small since energetic photons can disintegrate other nuclei and result in an inconsistency with observed abundances . in the case of low energy photons ,
abundances of only d , @xmath6li , and @xmath6be can be significantly affected because of their small threshold energies for photodisintegration @xcite . if the photon energy is larger ( 4 mev @xmath367 mev ) , the photodisintegration of @xmath178li , @xmath13he and @xmath13h , and the production of @xmath178li via @xmath6be(@xmath11 , @xmath179)@xmath178li and @xmath6li(@xmath11 , @xmath177)@xmath178li are also possible .
in addition , the time evolution of the baryon - to - photon ratio induced by the @xmath1 decay changes light element abundances from the values in sbbn .
therefore , we adopt the following constraints for respective light nuclides . it is noted that only the constraints on d and @xmath6li abundances are important while other constraints are not in deriving constraints in the parameter plane of this model with the decaying sterile neutrino of @xmath368 mev ( sec .
[ sec6b ] ) .
we use the primordial d abundance determined from observations of quasistellar object ( qso ) absorption systems . as conservative constraints ,
the @xmath344 and @xmath369 ranges estimated with the mean value of ten lyman-@xmath12 absorption systems , log(d / h)@xmath370 ( @xmath371 ) @xcite , are adopted .
@xmath13he abundances are measured in galactic regions through the @xmath13he@xmath372 8.665 ghz hyperfine transition line , @xmath13he / h=@xmath373 ( @xmath371 ) @xcite .
since the uncertainty in the estimation of primordial @xmath13he abundance is large , this constraint should be regarded only as a guide .
we take a @xmath344 upper limit from this observation .
primordial @xmath34he abundance has been derived by two different observations of metal - poor extragalactic regions : @xmath374 ( @xmath371 ) @xcite and @xmath375 ( @xmath371 ) @xcite .
we take @xmath344 limits from the latter conservative result with a large error bar .
primordial @xmath6li abundance is taken from a determination by spectroscopic observations of mpss .
the observed abundances are about three times smaller than theoretical values in the sbbn model .
we adopt the observational limit , log(@xmath6li / h)@xmath22 ( @xmath371 ) derived in a 3d nonlocal thermal equilibrium model @xcite .
we should consider the possibility that @xmath6li abundances in surfaces of mpss are depleted by a factor of @xmath376 as suggested by calculations of a stellar model with a turbulent mixing @xcite .
below we find that the @xmath6li abundance in this model can not be consistent with the observational @xmath344 limit .
when the @xmath6li depletion occurs in the stellar surfaces , however , some degree of @xmath6li reduction in this model can explain the observation .
@xmath178li abundances are measured in observations of mpss . a recent analysis , however
, does not indicate any detection of this isotope @xcite .
since @xmath178li nuclei can be reasonably produced and destroyed after bbn , its primordial abundance is chosen conservatively .
we use the least stringent @xmath344 upper limit among those for all stars reported in ref .
@xcite : @xmath178li / h=@xmath377 ( @xmath371 ) for the g64 - 12 ( nonlocal thermal equilibrium model with 5 free parameters ) . the effective neutrino number has been constrained from cmb observations .
we adopt the latest limit derived from the combined data of planck+wp+highl+bao for the one - parameter extension to the base @xmath51cdm model : @xmath378 ( 95% c.l . )
this one - parameter extension model is different from the present model of decaying sterile neutrino in terms of cosmic expansion rates and the time evolutions of @xmath5 values . in the former model , relativistic particles
are added to the standard @xmath51cdm model , while in the latter , the nonrelativistic @xmath1 and its decay are assumed .
the above limit , therefore , can not be applied to the present model as it is .
this limit , however , mainly comes from effects of the radiation energy density around the matter - radiation equal time .
since the equal time is much later than the decay lifetime of the sterile neutrino @xmath379 s ( see sec .
[ sec3 ] ) considered in this paper , the above limit approximately gives a limit on the final @xmath5 value in this decaying @xmath1 model .
an injection of nonthermal photons to the thermal bath triggers a deformation of the cbr spectrum from black - body @xcite .
such a deformation is severely constrained by observations which indicate a nearly complete planck spectrum @xcite .
we adopt the most stringent and reliable limits on the cmb energy spectrum as follows .
the limit on the chemical potential is taken from the analysis of the data from the far - infrared absolute spectrophotometer on board the cosmic background explorer , @xmath380 @xcite .
the limit on the compton @xmath381-parameter is taken from an updated constraint from the second generation of the absolute radiometer for cosmology , astrophysics , and diffuse emission ( arcade ) utilizing a better fitting procedure , @xmath382 @xcite .
figure [ fig4 ] shows normalized energy spectra of electron [ eq .
( [ eq_a23 ] ) ] and positron [ eq . ( [ eq_a24 ] ) ]
generated at the @xmath1 decay , and the total spectrum ( multiplied by 1/2 ) as a function of @xmath383 with @xmath86 the energies of electron and positron .
for all figures in this subsection , we assume that the mass of the sterile neutrino is @xmath384 mev since we find that this mass is included in the most important mass region for @xmath6be photodisintegration in the present model ( see sec .
[ sec6b ] ) . functions @xmath385 , @xmath386 , and @xmath387 [ cf .
( [ eq_a25])-([eq_a27 ] ) ] are also shown .
because of these extended energy spectra of @xmath3 , nonthermal photons produced via inverse compton scattering of cbr by the generated @xmath3 also have extended energy spectra . ) ] and positron [ eq .
( [ eq_a24 ] ) ] generated at the @xmath1 decay , and the total spectrum . functions @xmath385 ,
@xmath386 and @xmath387 [ cf .
( [ eq_a25])-([eq_a27 ] ) ] are also shown .
the mass of the sterile neutrino is assumed to be @xmath384 mev .
[ fig4],width=302 ] figure [ fig5 ] shows energy spectra of primary photon produced via the inverse compton scattering of electron and positron which are generated at the decay for the cosmic photon temperature of @xmath388 , 10 and 100 kev , respectively .
the primary photon spectra is given by @xmath389 when the energy injection with given spectra ( fig .
[ fig4 ] ) occurs at lower temperatures , the inverse compton scattering produces softer spectra of nonthermal primary photon , and cutoff energies are lower [ cf .
( [ e_g0_max ] ) ; positions of cutoff are not seen in this figure ] .
these results reflect the differential inverse compton scattering rate [ eq .
( [ eq_p_ic ] ) ] . , 10 and 100 kev , respectively .
the mass of the sterile neutrino is assumed to be @xmath384 mev .
[ fig5],width=302 ] figure [ fig6 ] shows injection spectra of nonthermal photon formed through the electromagnetic cascade calculated with eq .
( [ eq_inj ] ) for @xmath388 , 3 and 10 kev , respectively .
as the temperature decreases , the primary photon spectrum @xmath390 becomes softer so that less photons have high energies enough to disintegrate @xmath6be .
the spectrum of photons secondarily induced by the primary photons , however , has an upper cutoff at @xmath128 @xcite which scales as inverse of the temperature .
therefore , the cutoff energy is larger at lower temperature .
the cutoff can be seen at @xmath391 mev for @xmath392 kev .
because of the combination of the softness of the primary photon spectrum and the cutoff energy in electromagnetic cascade shower , there is a best temperature of the energy injection where relatively large abundances of energetic photons are produced with @xmath393 mev which can destroy @xmath6be .
one can find that the best temperature is 3 kev among the three temperatures shown in this figure . , 3 and 10 kev , respectively .
the mass of the sterile neutrino is assumed to be @xmath384 mev .
[ fig6],width=302 ] figure [ fig1 ] shows contours for calculated abundances of d and @xmath6li in the ( @xmath77 , @xmath198 ) plane for @xmath384 mev .
the two curved diagonal solid lines marked as `` d low ( @xmath344 ) '' and `` ( @xmath369 ) '' correspond to the observational @xmath344 and @xmath369 limits , respectively , on the d abundance .
the regions above the lines are excluded by d underproduction .
solid sharp curves at @xmath394 s are contours for the reduction ratio of @xmath6li abundance defined as @xmath395 where ( @xmath6li / h ) is the abundance calculated in this model , and ( @xmath6li / h)@xmath396 is the sbbn value . dashed curves are , on the other hand , contours for the calculated abundances ( @xmath6li / h ) . inside the solid curves for @xmath6li ,
the calculated @xmath6li abundance is smaller than the sbbn value because of the @xmath6be photodisintegration .
the @xmath6li abundance is larger above the dashed curves than the sbbn value because of larger @xmath4 value during the bbn epoch .
li in the parameter plane of ( @xmath77 , @xmath198 ) for @xmath384 mev .
the regions above the curved diagonal solid lines marked as `` d low ( @xmath344 ) '' and `` ( @xmath369 ) '' are excluded by d underproduction compared to the observational constraints at @xmath344 and @xmath369 levels , respectively .
the calculated @xmath6li abundance is smaller than the sbbn value above solid sharp curves at @xmath394 s by percentages shown near the curves , while it is larger above the dashed curves .
the dotted lines show the @xmath344 and @xmath369 upper limits on the effective neutrino number at the cosmological recombination epoch from cmb power spectrum .
the right region from the nearly - vertical solid line labeled as `` cmb '' is excluded from the limits from the cmb energy spectrum .
the shaded region corresponds to thermal freeze - out @xmath1 abundances with a possible @xmath1 dilution by a factor of 100 taken into account ( see sec .
[ sec7a ] and [ sec7b]).,width=302 ] the dotted lines show the @xmath344 and @xmath369 upper limits on the effective neutrino number at the cosmological recombination epoch from cmb power spectrum .
the nearly - vertical solid line labeled as `` cmb '' corresponds to the limits on the cmb energy spectrum .
the right region from this line ( long lifetime @xmath397 s ) is excluded from a large deformation in the energy spectrum .
we note that in the parameter region shown in fig .
[ fig1 ] , this model is constrained exclusively from the limit on the cmb @xmath217 parameter .
figure [ fig_neff ] shows contours of the ratio between the baryon - to - photon ratios in the bbn epoch and the cosmological recombination epoch ( solid lines ) , and contours of the @xmath5 value in the cosmological recombination epoch after the sterile neutrino decay in the parameter plane of ( @xmath77 , @xmath198 ) for @xmath384 mev .
value in the cosmological recombination epoch after the sterile neutrino decay in the parameter plane of ( @xmath77 , @xmath198 ) for @xmath384 mev .
[ fig_neff],width=302 ] the parameter region for small @xmath6li abundance is found at @xmath398 s and @xmath399 gev . in the epoch between the bbn and the matter radiation equality , the photon temperature in the universe @xmath81
is related to the cosmic age @xmath400 as @xmath401 kev @xmath402 unless the cosmic expansion rate is significantly affected by the sterile neutrino .
if the @xmath1 decay occurs at @xmath403 ( 1 kev ) , the @xmath6li abundance is reduced most effectively although some amount of d destruction always occurs simultaneously .
the shapes of contours of d and @xmath6li are explained as follows . at short lifetimes of @xmath404
s , the upper cutoff of the nonthermal photon spectrum @xmath128 @xcite is smaller than threshold energies for photodisintegration of light nuclides @xmath405 mev .
effects of nonthermal photons on elemental abundances are , therefore , negligibly small . at longer lifetimes of @xmath406
s , on the other hand , primary photons produced via the inverse compton scattering have softer energy spectra and lower cutoff energy originating from the maximum energy of the scattered photon [ eq .
( [ e_g0_max ] ) ] . this fact results in smaller abundances of nonthermal photons which are energetic enough to destroy d and @xmath6li . as a result ,
effects of nuclear photodisintegration are less efficient in the long lifetime region also .
the lower solid line of d is located between the solid curves of @xmath407li@xmath408 and @xmath409 in the interesting parameter region .
it means that in this region , the @xmath6be photodisintegration can slightly reduce the primordial @xmath6li abundance down to nearly the observational @xmath344 upper limit multiplied by the stellar depletion factor of two .
the d abundance , however , simultaneously decreases down to the observational @xmath344 lower limit .
we then find that there is no parameter region in which primordial @xmath6li abundance can be consistent with the mps value without assuming a stellar depletion .
this model , however , provides a mechanism of @xmath6li reduction by some small factor with its signatures imprinted in the primordial d abundance and the effective neutrino number ( sec .
[ sec6d ] ) .
figures [ fig2 ] and [ fig3 ] show contours for calculated abundances of d and @xmath6li , the effective neutrino number , and the cmb @xmath217 parameter as in fig .
[ fig1 ] , but for @xmath410 and @xmath411 mev , respectively .
as seen in figures [ fig1 ] , [ fig2 ] , and [ fig3 ] , regions of d and @xmath6li destruction at @xmath394 s are located at lower positions for larger masses .
this is because the sterile neutrinos with smaller masses can generate smaller numbers of energetic @xmath3 s which trigger d and @xmath6li destruction .
since the energy fraction of energetic photons capable of destroying @xmath6be to total nonthermal photons is smaller , we need a larger total energy generated at the @xmath1 decay . such a large energy injection is , however , constrained from bbn results taking into account the change of the baryon - to - photon ratio from the cmb - inferred value ( sec .
[ sec4c ] ) .
therefore , for @xmath412 mev the photodisintegration is less effective than the change of the baryon - to - photon ratio , and no parameter region for @xmath6li reduction is found .
the mass of @xmath413 mev is thus the best case for @xmath6be reduction that is a partial solution to the li problem .
but for @xmath410 mev .
[ fig2],width=302 ] but for @xmath414 mev .
[ fig3],width=302 ] the parameter region for the small primordial @xmath6li abundance is about to be excluded from observational constraint on d abundance . in this parameter region ,
the change of d abundance partially originates from the change of the baryon - to - photon ratio . in the case of smaller @xmath76 values , the change of the baryon - to - photon ratio is more significant . in the interesting parameter region ,
the effective neutrino number is also affected .
therefore , this model for the @xmath6li reduction will be tested through observational determinations of the effective neutrino number ( sec .
[ sec4b ] ) in near future .
we compare results of nonthermal nucleosynthesis calculated with the new and old cross sections of @xmath6be(@xmath11 , @xmath12)@xmath13he and @xmath6li(@xmath11 , @xmath12)@xmath13h .
old cross sections have been adopted from the fitted functions in ref .
for the reaction @xmath6be(@xmath11 , @xmath12)@xmath13he , the cross section is set to be zero for @xmath415 mev@xmath416 since the function gives negative values .
figure [ g_func ] shows the transfer functions @xmath195_a}^1(t ; 14~{\rm mev})$ ] [ eq .
( [ nonthermal_rate1b ] ) ] calculated with the new ( thick lines ) and old ( thin lines ) cross sections as a function of @xmath417 . the new transfer function of @xmath6be(@xmath11
, @xmath12)@xmath13he is 2.3 times smaller than the old one , while that of @xmath6li(@xmath11 , @xmath12)@xmath13h is 2.5 times smaller than the old one at their peak positions .
these changes are caused mainly by the fact that the new cross sections at low energies near the photodisintegration threshold energies are about one third of the old ones ( fig .
[ fig_a1 ] ) . _
a}^1(t ; 14~{\rm mev})$ ] [ eq .
( [ nonthermal_rate1b ] ) ] calculated with the new ( thick lines ) and old ( thin lines ) cross sections of @xmath6be(@xmath11 , @xmath12)@xmath13he and @xmath6li(@xmath11 , @xmath12)@xmath13h as a function of @xmath417 .
[ g_func],width=302 ] figure [ fig1_old ] shows the same contours for calculated abundances of d and @xmath6li , the effective neutrino number , and the cmb @xmath217 parameter as in fig .
[ fig1 ] for @xmath384 mev , but for results calculated with old cross sections .
solid lines for @xmath6li are located in the lower positions than those in fig .
this is because the adopted cross sections of @xmath6be(@xmath11 , @xmath12)@xmath13he and @xmath6li(@xmath11 , @xmath12)@xmath13h are smaller , and resultingly , the photodisintegration rates are smaller ( fig .
[ g_func ] ) .
we note that when old cross sections are used carelessly , one finds a fake parameter region in which @xmath6li abundance can be significantly reduced without a large efficiency of d photodisintegration , as shown in this figure .
it is , therefore , necessary to adopt the precise cross sections of the reactions @xmath6be(@xmath11 , @xmath12)@xmath13he and @xmath6li(@xmath11 , @xmath12)@xmath13h . for @xmath384 mev ,
but the old cross sections of @xmath6be(@xmath11 , @xmath12)@xmath13he and @xmath6li(@xmath11 , @xmath12)@xmath13h @xcite are used in the calculation .
[ fig1_old],width=302 ] figure [ fig_li7 ] shows contours of the ratio between the @xmath6li / h abundances calculated with the new and old cross sections of @xmath6be(@xmath11 , @xmath12)@xmath13he and @xmath6li(@xmath11 , @xmath12)@xmath13h in the parameter plane of ( @xmath77 , @xmath198 ) for @xmath384 mev . in the parameter region at @xmath418 s , the photodisintegration of @xmath6be and @xmath6li is efficient .
since the photodisintegration rates from the new cross sections are smaller than those from the old ones , larger abundances of @xmath6be and @xmath6li survive the photodisintegration .
the final abundance of @xmath6li , given by the sum of the @xmath6be and @xmath6li abundances , is , therefore , larger in the case of the new cross sections .
li / h abundances calculated with the new and old cross sections of @xmath6be(@xmath11 , @xmath12)@xmath13he and @xmath6li(@xmath11 , @xmath12)@xmath13h in the parameter plane of ( @xmath77 , @xmath198 ) for @xmath384 mev . [ fig_li7],width=302 ] we take a parameter set for the @xmath6be destruction : ( @xmath77 , @xmath198)=(@xmath419 s , @xmath420 gev ) in the case of @xmath384 mev . figure [ fig_evolution ] shows calculated nuclear abundances ( the top panel ) , the baryon - to - photon ratio ( the middle panel ) , and the effective neutrino number [ eq .
( [ eq_neff ] ) ] ( the bottom panel ) as a function of @xmath417 .
the solid and dotted lines correspond to results of the present model with the decaying sterile neutrino and the sbbn , respectively . in the top panel ,
@xmath421 and @xmath68 are the mass fractions of @xmath67h and @xmath34he , respectively , while other curves are number densities of other nuclides relative to that of hydrogen . ) ] ( the bottom panel ) as a function of @xmath417 .
the solid and dotted lines show results of this model with the decaying sterile neutrino and the sbbn , respectively . in the top panel ,
@xmath421 and @xmath68 are the mass fractions of @xmath67h and @xmath34he , respectively , while abundances of other nuclides are given by ratios of number densities of nuclides and hydrogen .
the mass , lifetime , and abundance of the sterile neutrino are set to be @xmath384 mev , @xmath422 s , and @xmath423 gev , respectively .
[ fig_evolution],width=302 ] in the top panel , effects of photodisintegration can be observed as differences of solid and dotted lines at @xmath424 .
the @xmath6be nuclei are disintegrated via @xmath6be(@xmath11 , @xmath12)@xmath13he , and the @xmath6be abundance is slightly decreases .
the deuterons are also disintegrated via @xmath425h(@xmath11 , @xmath177)@xmath67h , and the d abundance decreases and the neutron abundance increases .
when the photodisintegration occurs , the temperature is already enough low that thermal nuclear reactions between charged nuclei are no longer operative .
however , nonradiative neutron capture reactions are operative since no coulomb repulsion exists in reactions with neutrons .
the generated neutrons are , therefore , partially captured by @xmath13he [ via @xmath13he(@xmath177 , @xmath179)@xmath13h ] and @xmath6be [ via @xmath13be(@xmath177 , @xmath179)@xmath6li ] . as a result , abundances of @xmath13h and @xmath6li slightly increase . in the middle panel , the baryon - to - photon ratio @xmath4 decreases at @xmath426 because of an electromagnetic energy injection at the @xmath1 decay .
since the final @xmath4 value is fixed to the planck value , the baryon - to - photon ratio in the @xmath1 model is higher than that in sbbn at @xmath427 .
although this difference of the @xmath4 value slightly changes bbn , differences in nuclear abundances during bbn epoch ( @xmath428 ) are so small that they can not be seen well .
the downturns of the solid and dotted lines at @xmath429 are caused by an entropy transfer from @xmath3 to photon at the cosmological @xmath3 annihilation .
the @xmath4 values before the @xmath3 annihilation are @xmath430 in the @xmath1 model and @xmath431 in the sbbn model , respectively . in the bottom panel ,
the solid and dashed lines descend at @xmath429 by the entropy transfer from @xmath3 to photon . the @xmath5 value
is then increased at @xmath426 by the nonthermal neutrino injection at the @xmath1 decay .
the final @xmath5 values after the @xmath1 decay are 3.19 in the @xmath1 model and 3 in the sbbn model , respectively .
we check the validity of the approximate formula for the final @xmath5 value .
the present lifetime corresponds to @xmath432 s ( sec .
[ sec4d2 ] ) and @xmath433 kev .
the present mass corresponds to parameters of @xmath434 and @xmath435 [ eq . ( [ eq_a12 ] ) ] .
the ratio of decay rates is @xmath436 [ eq . ( [ eq_r_dec ] ) ] .
the ratios of the average energies to the sterile neutrino mass in the decay mode @xmath287 are @xmath437 , @xmath438 , and @xmath439 .
the ratio of injected energies of neutrinos and @xmath3 is then derived as @xmath440 [ eq .
( [ eq_ratio_nue ] ) ] . the corresponding ratio of entropy densities before and after the @xmath1 decay is @xmath441 [ eq . ( [ eq_entropy ] ) ] .
the effective neutrino number contributed from nonthermal neutrinos without the entropy production effect is derived as @xmath442 [ eqs .
( [ eq_zeta_ratio ] ) and ( [ neff_nt_i ] ) ] .
taking account of the dilution associated with the entropy production , the final effective number is @xmath443 [ eq .
( [ eq_dn ] ) ] .
we note that a slight change in the cosmic expansion rate through a change in the @xmath444 value by the @xmath1 was neglected in this estimation .
it is thus found that this approximate estimation successfully gives the final @xmath5 value derived in our precise numerical calculation .
it is assumed that the sterile neutrino has a mass @xmath76 after the ew phase transition .
depending on the mixing angle , the sterile neutrino react with standard model particles mainly via the following weak reactions : @xmath445 where @xmath446 is any fermion , i.e. , charged leptons @xmath447 [ @xmath266 ( @xmath448 ) , @xmath449 ( @xmath450 ) , and @xmath451 ( @xmath452 ) ] , neutrinos @xmath453 and up - type quarks @xmath454 [ @xmath455 ( @xmath448 ) , @xmath38 ( @xmath450 ) , and @xmath400 ( @xmath452 ) ] , and down - type quarks @xmath456 [ @xmath174 ( @xmath448 ) , @xmath251 ( @xmath450 ) , and @xmath42 ( @xmath452 ) ] . in the charged current reactions , probabilities of producing respective flavors ( @xmath457 )
are described by the cabibbo - kobayashi - maskawa matrix @xcite . when the weak interaction rate becomes smaller than the hubble expansion rate , the abundance of the sterile neutrino freezes out from equilibrium .
thereafter , the ratio between the @xmath1 number density and the entropy density @xmath458 does not change . is measured in a unit different from that of nuclear mole fractions @xmath459 introduced in sec .
[ sec2 g ] .
] weak interaction rates of sterile neutrinos , @xmath1 s , with weakly interacting standard model particles after the ew phase transition scale @xcite as @xmath460 where @xmath461 is the fermi constant , @xmath462 is the mixing angle . the ratio between the rates and cosmic expansion rate , @xmath96 ,
is then given @xcite by @xmath463{\mathstrut 5 } } \frac{g_\ast^{1/2 } t^2}{m_{\rm pl } } \right)^{-1 } \nonumber\\ & = & \frac{3 \sqrt[]{\mathstrut 5}}{2\pi^{3/2 } } \frac{m_{\rm pl}g_{\rm f}^2 \theta^2 t^3}{g_\ast^{1/2 } } \nonumber\\ & = & 9.69 \times 10^{7 } \left(\frac{\theta}{10^{-3 } } \right)^2 \left(\frac{g_\ast}{106.75}\right)^{-1/2 } \left(\frac{t}{100~{\rm gev}}\right)^3 \nonumber\\ & = & 1.00 \left(\frac{\theta}{10^{-3 } } \right)^2 \left(\frac{g_\ast } { 63.75}\right)^{-1/2 } \left(\frac{t}{0.2~{\rm gev}}\right)^3 .
\label{eq_temp_freezeout } \ ] ] in the last line of this equation , the statistical dof of the sterile neutrino , i.e. , @xmath464 were added to the value of @xmath465 at @xmath466 mev in the standard model .
a sterile neutrino with mixing angle @xmath467 would thus freeze out from equilibrium at temperature @xmath468 mev .
the relic abundance of @xmath1 is , therefore , given by the abundance fixed at @xmath468 mev .
the lifetime of the sterile neutrinos is roughly given [ cf .
( [ eq_a9 ] ) and ( [ eq_a21 ] ) ] by @xmath469 therefore , if a sterile neutrino with a mass @xmath413 mev decays @xmath470 s after the big bang and reduces the primordial @xmath6li abundance , the mixing angle would be @xmath471 .
the time evolution of the @xmath1 abundance has been calculated @xcite . in refs .
@xcite , however , the maximal mixing angle @xmath472 is implicitly assumed , and the dependence on the mixing angle @xmath91 is not considered .
in addition , the authors took into account only the annihilation @xmath473 @xcite , which is negligibly weaker than the reactions eqs .
( [ process_1])-([process_9 ] ) when @xmath474 . in ref .
@xcite , on the other hand , a dedicated calculation has been performed . however , the authors only focused on shorter @xmath1 lifetimes of @xmath475 s , which correspond to relatively large values of the mixing angle @xmath476 compared with those considered in this paper . depending on the mixing angle , the weak reaction freeze - out of the sterile neutrino occurs in various epochs with different values of @xmath444 .
perhaps the sterile neutrino never experiences the weak reaction equilibrium after the ew phase transition . in general
, the @xmath1 relic abundance can sensitively depend on the evolution of sterile neutrino mass during the ew phase transition , which differs in different models of the sterile neutrino .
precise calculations of the @xmath1 relic abundance should be performed in detail .
however , they are beyond the scope of this paper .
we estimate the freeze - out abundance of the sterile neutrino as a function of @xmath76 and @xmath77 as follows . for a given set of ( @xmath76 , @xmath77 ) ,
a corresponding @xmath91 value is derived with eq .
( [ gamma_approx ] ) . the temperature satisfying eq .
( [ eq_temp_freezeout ] ) is then derived with the @xmath91 value .
this temperature is defined as the freeze - out temperature @xmath477 .
an approximate value of the freeze - out abundance of @xmath1 is given by the equilibrium abundance at @xmath477 .
for the @xmath478 value , the freeze - out abundance is given by the equilibrium abundance @xmath479 using the following equations .
the equilibrium number density of a fermion is given @xcite by @xmath480 where @xmath481 and @xmath482 are the mass and statistical dof , respectively , of the fermion @xmath18 , and @xmath483 is a function given by @xmath484 in the nonrelativistic limit , the function @xmath485 has the limit value of @xmath486 .
the entropy density is given @xcite by @xmath487 from eqs .
( [ n_eq ] ) and ( [ h_plus ] ) , the abundance ratio is given by @xmath488 figure [ fig_relic ] shows massless dofs in terms of energy and entropy , i.e. , @xmath444 and @xmath336 , respectively , as a function of photon temperature @xmath81 .
solid lines for massless dofs correspond to the standard model plus a sterile neutrino of mass @xmath384 mev and statistical dof of @xmath489 , while dashed lines correspond to the standard model .
also shown is the equilibrium abundance ratio of a sterile neutrino @xmath490 in the model with the sterile neutrino .
the massless dofs are calculated as in ref .
@xcite based on the latest data on particle mass @xcite .
it is assumed that the quark hadron transition occurs suddenly at temperature @xmath491 mev .
above the temperature , quarks are taken into account in the dofs .
below the temperature , on the other hand , contributions of only hadrons are included and those of quarks are neglected .
we only take into account dofs of charged and neutral pions at @xmath492 since they are only relativistic hadrons . and
@xmath336 , respectively , as a function of photon temperature @xmath81 .
solid lines for massless dofs correspond to the standard model plus a sterile neutrino of mass @xmath384 mev and statistical dof of @xmath489 , while dashed lines correspond to the standard model .
also shown is the equilibrium abundance ratio of a sterile neutrino @xmath490 in the model with the sterile neutrino .
[ fig_relic],width=302 ] the massless dofs in the model with @xmath1 is larger than those in the model without @xmath1 by about two because of the statistical dof of sterile neutrino .
as the temperature decreases , weak bosons , heavy quarks and leptons become nonrelativistic , and the dofs become small . at the quark hadron transition temperature @xmath493 , dofs of quarks and gluons disappear , and the dofs drastically decreases .
the equilibrium abundance @xmath490 increases as the temperature increases since it is proportional to the @xmath494 value . at @xmath495 ,
the abundance significantly increases . at the lowest temperature of @xmath496 mev ,
the sterile neutrino start to be nonrelativistic .
the equilibrium abundance then decreases from this temperature .
if a light sterile neutrino with the mass @xmath497 mev survives during bbn epoch , its number density must have diluted between its weak freeze - out [ @xmath498 mev ] and the bbn epoch [ @xmath499 mev ] in order to avoid a large change of the baryon - to - photon ratio associated with the @xmath1 decay ( see sec .
[ sec7b ] ) .
for example , we consider the case of @xmath384 mev and @xmath423 gev .
this assumption corresponds to the energy ratio @xmath500 [ eqs .
( [ eq_ratio_nue ] ) and ( [ eq_zeta_ratio ] ) ] , and the total energy injection of @xmath501 gev .
this energy injection is realized by the decay of sterile neutrino with the mass @xmath384 mev and the number ratio @xmath502 , where we used a relation , @xmath503 where the ratio @xmath504 should be measured after the cosmological @xmath3 annihilation .
however , the freeze - out abundance is @xmath505 .
therefore , the sterile neutrino needs to be diluted by a factor of several hundreds .
it is shown that the decays of heavier sterile neutrinos into standard model particles can realize a dilution of the light sterile neutrino ( @xmath506 mev ) although our naive estimate @xcite suggests that a decay of a heavier sterile neutrino would result in a dilution factor smaller than required for the appropriate abundance by some factor at least .
we assume that one of heavier sterile neutrinos , i.e. , @xmath507 , predominantly contributes to the dilution or entropy production .
in addition , it is assumed that the heavy neutrino dominates in terms of energy density in its decay epoch . supposing that @xmath507 decays into relativistic leptons and quarks which are thermalized rapidly with respect to the cosmic expansion time scale , the energy density of relativistic species after the decay is @xmath508 , where @xmath509 is the reheating temperature .
this energy density is equal to the energy density of @xmath507 before the decay . the ratio of the entropy per comoving volume at the epoch long after the decay ( aft ) to that long before the decay ( bef ) is given @xcite by @xmath510 where @xmath511 and @xmath512 ( for @xmath513bef and aft ) are the scale factor and the photon temperature of the universe at time @xmath514 , and we supposed @xmath515 and that the @xmath444 and @xmath336 values do not change between the temperatures of @xmath516 and @xmath517 .
we note that a large dilution factor is realized only for @xmath518 gev .
if the mass is much larger than the energy scale of the ew phase transition , the freeze - out @xmath507 abundance is small because of the boltzmann suppression factor .
furthermore , the lifetime should not be longer than @xmath519 s since bbn is significantly affected by nonthermal reactions of hadronic particles generated at the @xmath507 decay if the lifetime is longer @xcite . from this equation , we find that the dilution factor , that equals the entropy enhancement factor , is about a factor of 100 at maximum .
this maximum factor is @xmath323 times smaller than the necessary factor of 300 .
some other mechanism of the dilution is , therefore , needed for the light sterile neutrino to destroy some moderate fraction of primordial @xmath6be successfully .
possible mechanisms include dilutions by massive particles other than @xmath507 decaying into active neutrinos , @xmath520 @xcite or photons @xmath521 ( @xmath522 ) @xcite . in figs .
[ fig1 ] , [ fig2 ] , [ fig3 ] , and [ fig1_old ] , parameter regions of thermal freeze - out @xmath1 abundances are shown by shaded regions .
a possible dilution of the sterile neutrino by a factor of 100 is taken into account .
the freeze - out abundances without dilutions are higher than the figure domains , and therefore not seen .
the lower boundaries of these regions correspond to the abundances diluted by a factor of 100 .
sudden drops of the boundaries at @xmath523 s result from the decrease in the massless dof in terms of entropy ( fig .
[ fig_relic ] ) .
it is clear that the parameter regions for the primordial @xmath6li reduction are lower than the regions of freeze - out abundances .
therefore , a dilution of the sterile neutrino is necessary for the @xmath6li reduction to work .
if either muon or tauon neutrino predominantly couples to @xmath1 and couplings of other charged leptons are negligible as an extreme case opposite to the case studied in this paper , effects of the sterile neutrino decay are changed .
for example , we take the case of @xmath384 mev .
the ratio of energy injections in the forms of @xmath3 and @xmath286 is @xmath524 for the coupling to @xmath525 , while it is @xmath526 for the coupling to @xmath527 ( @xmath528 or @xmath197 ) [ eq .
( [ eq_zeta_ratio ] ) ] .
the muon or tauon type mixings , therefore , result in a large energy fraction of neutrino emitted at the decay .
an @xmath3 injection decreases the @xmath4 and @xmath5 values and induces nonthermal nucleosynthesis , while a neutrino injection increases the @xmath5 value .
therefore , a sterile neutrino that mixes only with @xmath527 has small effects on the primordial light element abundances and the @xmath4 value relative to that on @xmath5 .
we assume that the sterile neutrino has a mass @xmath529 mev and a lifetime @xmath530 s ( parameter value for the @xmath6li reduction ) , and that mixing angles of muon and tauon types , @xmath531 and @xmath532 , respectively , can be neglected .
the active - sterile mixing angle is then determined to be @xmath533 [ eqs .
( [ gamma_approx ] ) and ( [ eq_a9 ] ) ] .
if the muon type mixing is sizable , we should take into account another constraint from low energy phenomena : the sterile neutrino can be produced by the decay of charged pions , e.g. , @xmath534 or @xmath535 .
this channel has been searched for a long time and many experiments gave constraints on the active - sterile mixing angle . from ref .
@xcite , @xmath536 should be smaller than @xmath537 . if the precision of those experiments can be improved by a few orders of magnitude , therefore , we may see a signal from pion decays , or exclude the possibility of primordial @xmath6li reduction by @xmath1 suggested in this paper . furthermore ,
if the muon type mixing is of the order of @xmath538 , the sterile neutrino might be detected by super kamiokande in future @xcite .
it is worth mentioning that a constraint from supernova sn1987a observation is rather strong @xcite ; that is roughly @xmath539 for any flavors .
the primordial lithium abundance determined from spectroscopic observations of mpss is smaller than the theoretical prediction of sbbn model by a factor of @xmath0 .
it has been suggested that a bbn model with a long - lived radiatively decaying exotic particle possibly provides a solution to the li problem . in that model
nonthermal photons with energies @xmath29 mev generated by the particle decay disintegrate @xmath6be . the primordial @xmath6li abundance , which is the sum of abundances of @xmath6li and @xmath6be produced during bbn ,
is then reduced . in this paper , we studied the possibility of @xmath2 mev sterile neutrino @xmath1 . if it decays after bbn , and electron and positron @xmath3 s
are emitted , the energetic @xmath3 s can produce energetic photons via the inverse compton scattering of cbr .
the solution to the li problem is , therefore , also expected in this model .
then , we investigated cosmological effects of the sterile neutrino decay on primordial light element abundances , the baryon - to - photon ratio , and the effective neutrino number .
the sterile neutrino was assumed to live longer than the bbn time scale , i.e. , @xmath540 s. this long lifetime satisfies a recent constraint from neutrino oscillation experiments : the lifetime of sterile neutrino lighter than the pion can not be shorter than @xmath33 s. we constructed a numerical code for consistent calculations of the cosmic expansion history , bbn , and nonthermal nucleosynthesis triggered by the @xmath1 decaying after bbn . the updated relation between the baryon - to - photon ratio and the baryon density of the universe is used ( sec .
[ sec2a1 ] ) .
the initial abundance , mass , and lifetime of the sterile neutrino were taken as free parameters .
then we formulated the injection spectrum of nonthermal photon induced by the @xmath1 decay ( sec .
[ sec2 ] ) .
we introduced an active - sterile mixing angle , and calculated the energy spectra of @xmath3 s and active neutrinos generated at the @xmath1 decay .
taking into account the primary photon production via the inverse compton scattering of cbr by energetic @xmath3 s , and electromagnetic cascade showers induced by the primary photons , the steady state injection spectrum was derived as a function of the sterile neutrino mass and the photon temperature .
nonthermal nucleosynthesis triggered by the energetic photons are then calculated . in this paper , we corrected errors in photodisintegration cross sections of @xmath6be(@xmath11 , @xmath12)@xmath13he and @xmath6li(@xmath11 , @xmath12)@xmath13h adopted in previous studies .
we gave functions for the cross sections in light of recent nuclear experimental results ( sec .
[ sec3 ] ) .
furthermore , effects of the @xmath1 decay on the cosmic thermal history and evolutions of effective neutrino number and the baryon - to - photon ratio are formulated exactly ( sec .
[ sec4 ] ) .
results of our calculations are summarized as follows : first , we calculated injection spectra of nonthermal photon as a function of the mass @xmath76 and photon temperature @xmath81 .
we took into account the energy spectra of @xmath3 s emitted at the decay , inverse compton scattering of cbr by the energetic @xmath3 s producing primary energetic photons , and electromagnetic cascade showers induced by the primary photons .
the energy spectra of @xmath3 s are broadly extended independent of the temperature .
the energy spectra of the primary photons , on the other hand , depend significantly on the temperature , and spectra are softer at lower temperatures in the later universe .
the final injection spectra of nonthermal photon also depend on the temperature significantly .
abundances of energetic photons capable of disintegrating @xmath6be are determined by the hardness of the primary photon spectra and the upper cutoff in the nonthermal photon spectra due to the double photon pair annihilation .
we found that an effective @xmath6be destruction can occur only if the sterile neutrino decays at @xmath7 kev ( sec .
[ sec6a ] ) .
second , we simultaneously solved nonthermal nucleosynthesis induced by the nonthermal photons , and evolutions of the baryon - to - photon ratio @xmath4 and the cosmological effective neutrino number @xmath5 . at the @xmath1 decay , energetic active neutrinos , electrons , and positrons
are generated .
the energies of the neutrinos are never thermalized since the weak interaction has been long since decoupled in the universe .
the nonthermal neutrinos , therefore , contribute to only the radiation energy density or the @xmath5 value .
the energies of the @xmath3 s are , on the other hand , quickly thermalized through interactions with cbr , and eventually transferred to cbr .
the comoving photon entropy is then increased .
using formulae relevant to the sterile neutrino decay ( appendix [ app1 ] ) , we quantitatively solved changes of @xmath4 and @xmath5 caused by the @xmath1 decay .
the final @xmath4 value is fixed to the planck value at the cosmological recombination .
calculated results are compared with observational constraints . as a result
, amounts of energy injection in the form of @xmath3 s at the @xmath1 decay are constrained from limits on primordial nuclear abundances ( d and @xmath6li ) , the effective neutrino number , and the cmb energy spectrum .
we found a parameter region of the lifetime @xmath77 and the amount of energy injection @xmath198 , in which @xmath6be is photodisintegrated and the li problem is partially solved : ( @xmath77 , @xmath198 ) = ( @xmath8 s , @xmath541 gev ) .
we also found that the sterile neutrino mass is required to be @xmath413 mev .
a lighter neutrino can not destroy any significant fraction of @xmath6be via photodisintegration without violating the constraints on the d abundance or the effective neutrino number .
the best parameter region is narrow even in the case of @xmath413 mev . in this parameter region ,
the @xmath6be destruction by more than a factor of three can not be realized since the constraint on the d abundance excludes this possibility ( sec .
[ sec6b ] ) .
third , it was found that in the best parameter region , the @xmath1 decay not only decreases the @xmath4 value slightly but also increases the @xmath5 value by a factor of @xmath10 1 . for the moment
, the 2 @xmath542 ranges of the d abundance from qso observations and the @xmath5 value from cmb observations do not indicate any effect by the sterile neutrino decay as considered in this paper .
the @xmath4 value at the cosmological recombination is consistent with the value at the bbn epoch inferred from measurements of primordial light element abundances , and the effective neutrino number is consistent with the case of only three active neutrino , i.e. , @xmath543 .
since error bars on the @xmath4 and @xmath5 values are getting smaller , this model for the li reduction can be tested by future observations of the parameters @xmath4 and @xmath5 ( sec . [ sec6b ] ) .
fourth , we compared results of the @xmath1 decay with the new and old cross sections of @xmath6be(@xmath11 , @xmath12)@xmath13he and @xmath6li(@xmath11 , @xmath12)@xmath13h .
the new rates for the former and latter reactions are 2.3 and 2.5 times , respectively , larger than the corresponding old rates .
the corrected cross sections thus resulted in significantly smaller efficiencies of @xmath6be and @xmath6li photodisintegration .
therefore , one should adopt the precise cross sections in calculations of nonthermal nucleosynthesis ( sec .
[ sec6c ] ) .
fifth , the thermal freeze - out abundance of the sterile neutrino was estimated and compared with the best parameter region for the @xmath6li reduction . the freeze - out abundance is much larger than the value required for the @xmath6li reduction .
therefore , the relic sterile neutrino must be diluted before the bbn epoch by some mechanism .
a sufficiently large dilution is , however , not realized by a decay of another sterile neutrino with a mass smaller than the ew scale of @xmath544 gev .
for example , therefore , other particles decaying before the bbn epoch are required for a successful @xmath6be destruction associated with the @xmath1 decay studied in this paper ( sec .
[ sec7 ] ) .
we are grateful to shintaro eijima and kazuhiro takeda for helpful comments .
this work was supported by inoue foundation for science [ hi ] .
we derive the total decay rate of sterile neutrinos , and energy spectra and average energies of electron and positron generated in the decay .
the overall amplitude of the matrix element squared for the decay of @xmath545 is given by @xmath546 \label{eq_a1}\ ] ] where @xmath461 is the fermi constant , @xmath474 is the mixing angle , @xmath547 is the four momentum of particle @xmath18 , the subscript @xmath18 identifies the particle species as @xmath548 for @xmath549 , 2 for @xmath527 , 3 for @xmath266 , and 4 for @xmath267 , and constant parameters @xmath40 , @xmath43 and @xmath172 are defined as @xmath550 where @xmath551 and @xmath552 are the constants for vector and axial couplings of charged leptons to the @xmath553 weak boson with @xmath554 @xcite the weak angle .
the @xmath40 term and the first terms of @xmath43 and @xmath172 correspond to the @xmath553 exchange , while the second term of @xmath43 corresponds to the @xmath555 exchange .
the third term of @xmath43 and the second term of @xmath172 correspond to the interference contribution .
when @xmath556 is satisfied , parameter values are @xmath557 , @xmath558 , and @xmath559 . the differential decay rate as a function of energies of @xmath266 and @xmath267 , i.e. , @xmath560 and @xmath561 ,
is then given by @xmath562 , \label{eq_a5}\ ] ] where new dimensionless variables were defined as follows : @xmath296 and @xmath563 .
the differential decay rates as a function of @xmath564 and @xmath565 are given @xcite by @xmath566\right\}_{x_{{\rm f}3}=x_{{\rm f}3,-}}^{x_{{\rm f}3,+ } } , \label{eq_a6 } \\
\frac { d \gamma}{dx_4 } & = & \frac{g_{\rm f}^2 \theta^2 m_{{\nu_{\rm h}}}^5}{64 \pi^3 } \left\ { a \left(\frac{x_{{\rm f}4}^2}{2}- \frac{x_{{\rm f}4}^3}{3 } \right ) + b x_4 \left(1 - x_4 \right ) x_{{\rm f}4 } + 2 c x_{\rm m}^2 \left[\left(2 - x_4\right ) x_{{\rm f}4 } -\frac{x_{{\rm f}4}^2}{2 } \right]\right\}_{x_{{\rm f}4}=x_{{\rm f}4,-}}^{x_{{\rm f}4,+ } } , \label{eq_a7}\end{aligned}\ ] ] where @xmath567 is the variable integrated in deriving eqs .
( [ eq_a6 ] ) and ( [ eq_a7 ] ) from eq .
( [ eq_a5 ] ) ( @xmath568 and @xmath569 ) , @xmath570 and @xmath571 are its minimum and maximum values , respectively , and the terms in braces are evaluated as differences between values for @xmath572 , i.e. , @xmath573 . these rates are derived by integration of eq .
( [ eq_a5 ] ) over @xmath574 in the range of @xmath575 . on the other hand , the ranges of @xmath564 and @xmath565 in eqs .
( [ eq_a6 ] ) and ( [ eq_a7 ] ) , respectively , are @xmath576 .
the values @xmath572 are given @xcite by @xmath577{\mathstrut x_i^2 - 4x_{\rm m}^2 } } { 2 \left(1 + x_{\rm m}^2 -x_i\right)}. \label{eq_a8}\ ] ] the total decay rate is given @xcite by @xmath581{\mathstrut 1- 4x_{\rm m}^2 } -12 x_{\rm m}^4 \left ( 1- x_{\rm m}^4 \right ) l \right ] \right .
\nonumber\\ & & \left .
~~~~~~~~~~~~~~~ + 4 c_2 \left [ x_{\rm m}^2 \left ( 2 + 10 x_{\rm m}^2 -12 x_{\rm m}^4 \right ) \sqrt[]{\mathstrut 1- 4x_{\rm m}^2 } + 6 x_{\rm m}^4 \left ( 1- 2x_{\rm m}^2 + 2x_{\rm m}^4 \right ) l \right ] \right\ } , \label{eq_a9}\end{aligned}\ ] ] where @xmath582{\mathstrut 1 - 4x_{\rm m}^2}}{x_{\rm m}^2 \left ( 1 + \sqrt[]{\mathstrut 1- 4x_{\rm m}^2 } \right ) } \right ] .
\label{eq_a12}\end{aligned}\ ] ] the adopted weak angle of @xmath554 @xcite corresponds to the values of @xmath583 and @xmath584 for the neutrino flavor of the final state @xmath556 .
the number spectra of the electron and positron emitted at the decay are given by @xmath585 the total spectra of electron and positron is given by @xmath586 + 2 c x_{\rm m}^2 x_{{\rm f } } \left(4 - 2x - x_{{\rm f } } \right ) \right\}_{x_{{\rm f}}=x_{{\rm f},-}}^{x_{{\rm f},+}}. \label{eq_a15}\end{aligned}\ ] ] the average energies of electron ( @xmath560 ) and positron ( @xmath561 ) are given by @xmath587 where we defined a function : @xmath588{\mathstrut 1- 4x_{\rm m}^2 } \left(3 - 29 x_{\rm m}^2 + 48x_{\rm m}^4 -70 x_{\rm m}^6 -60 x_{\rm m}^8 \right ) \right .
\nonumber\\ & & \left .
~~~~~ - x_{\rm m}^4 \left [ \left(1+x_{\rm m}^2\right ) \left ( 1-x_{\rm m}^4 \right ) l_1 + \left ( 3- x_{\rm m}^2 + x_{\rm m}^4 + x_{\rm m}^6 \right ) l_2 \right ] \right\ } \nonumber\\ & & -b \left\ { \frac{1}{60 } \sqrt[]{\mathstrut 1- 4x_{\rm m}^2 } \left(9 - 52 x_{\rm m}^2 + 14x_{\rm m}^4 + 80 x_{\rm m}^6 + 120 x_{\rm m}^8 \right ) \right .
\nonumber \\ & & \left .
~~~~~~~ + x_{\rm m}^4 \left [ \left(1-x_{\rm m}^4 -2x_{\rm m}^6 \right ) l_1 + \left ( 3 + x_{\rm m}^4 + 2x_{\rm m}^6 \right ) l_2 \right ] \right\ } \nonumber \\ & & + c \left\ { \frac{1}{6 } x_{\rm m}^2 \sqrt[]{\mathstrut 1- 4x_{\rm m}^2 } \left(5 - 12 x_{\rm m}^2 + 10x_{\rm m}^4 -12 x_{\rm m}^6 \right ) \right . \nonumber \\ & & \left .
~~~~~~~ + 2x_{\rm m}^4 \left [ \left(1-x_{\rm m}^2\right ) \left ( 1-x_{\rm m}^4 \right ) l_1 + \left ( 1 + x_{\rm m}^2 + x_{\rm m}^4 -x_{\rm m}^6 \right ) l_2 \right ] \right\ } \nonumber \\ & & + ( b - 2 c x_{\rm m}^2 ) \left\ { \frac{1}{24 } \sqrt[]{\mathstrut 1- 4x_{\rm m}^2 } \left(5 -38x_{\rm m}^2 + 6x_{\rm m}^4 + 36 x_{\rm m}^6 \right ) \right .
\nonumber \\ & & \left .
~~~~~~~~~~~~~~~~~~~~ - \frac{1}{2 } x_{\rm m}^4 \left(1 + 3x_{\rm m}^4 \right ) \left(l_1 -l_2
\right ) \right\ } , \label{eq_a18}\end{aligned}\ ] ] where parameters @xmath589 and @xmath590 are defined as @xmath591{\mathstrut 1- 4x_{\rm m}^2}}{2x_{\rm m}^3 } \right ] , \label{eq_a19}\\ l_2 & = & \ln \left [ \frac{1 + \sqrt[]{\mathstrut 1- 4x_{\rm m}^2}}{2x_{\rm m } } \right ] , \label{eq_a20}\end{aligned}\ ] ] and @xmath592 is satisfied .
figure [ fig7 ] shows average energies of electron , positron and neutrino generated at the @xmath1 decay as a function of @xmath299 calculated with eqs .
( [ eq_a16 ] ) and ( [ eq_a17 ] ) and a trivial relation of @xmath593 . in the small @xmath299 region ,
average energies of all the three particles in the final state are close to one third of the sterile neutrino mass @xmath76 , i.e. , @xmath594 . in the large @xmath595 region ,
on the other hand , masses of electron and positron are significant fractions of the sterile neutrino mass .
the most of the energy in the final state is , therefore , taken for the mass energy , and the average energy of @xmath525 is small . finally , the decay rate for the mode of @xmath596 is given @xcite by @xmath597 we note that the decay into the final state of @xmath598 for @xmath528 and @xmath197 does not occur in the assumption adopted in this paper ( see sec . [ sec2c ] ) .
a. coc , s. goriely , y. xu , m. saimpert and e. vangioni , astrophys . j. * 744 * , 158 ( 2012 ) .
a. coc , j. -p .
uzan and e. vangioni , arxiv:1307.6955 [ astro-ph.co ] .
d. n. spergel _ et al . _
[ wmap collaboration ] , astrophys .
j. suppl .
* 148 * , 175 ( 2003 ) d. n. spergel _ et al . _
[ wmap collaboration ] , astrophys .
j. suppl .
* 170 * , 377 ( 2007 ) .
d. larson _
et al . _ ,
astrophys .
j. suppl . * 192 * , 16 ( 2011 ) .
g. hinshaw , d. larson , e. komatsu , d. n. spergel , c. l. bennett , j. dunkley , m. r. nolta and m. halpern _ et al .
_ , astrophys .
j. suppl .
* 208 * , 19 ( 2013 ) .
p. a. r. ade _ et al .
_ [ planck collaboration ] , arxiv:1303.5076 [ astro-ph.co ] .
f. spite and m. spite , astron .
astrophys .
* 115 * , 357 ( 1982 ) .
s. g. ryan , t. c. beers , k. a. olive , b. d. fields and j. e. norris , astrophys . j. * 530 * , l57 ( 2000 ) .
j. melendez and i. ramirez , astrophys . j. * 615 * , l33 ( 2004 ) .
m. asplund , d. l. lambert , p. e. nissen , f. primas and v. v. smith , astrophys .
j. * 644 * , 229 ( 2006 ) .
p. bonifacio _ et al .
_ , astron .
astrophys .
* 462 * , 851 ( 2007 ) .
w. aoki , p. s. barklem , t. c. beers , n. christlieb , s. inoue , a. e. g. perez , j. e. norris and d. carollo , astrophys .
j. * 698 * , 1803 ( 2009 ) .
j. i. g. hernandez , p. bonifacio , e. caffau , m. steffen , h. -g .
ludwig , n. t. behara , l. sbordone and r. cayrel _ et al .
_ , astron . astrophys . *
505 * , l13 ( 2009 ) .
l. sbordone , p. bonifacio , e. caffau , h. -g .
ludwig , n. t. behara , j. i. g. hernandez , m. steffen and r. cayrel _ et al .
_ , astron .
astrophys . * 522 * , a26 ( 2010 ) .
l. monaco , s. villanova , p. bonifacio , e. caffau , d. geisler , g. marconi , y. momany and h. -g .
ludwig , astron .
astrophys .
* 539 * , a157 ( 2012 ) .
a. mucciarelli , m. salaris and p. bonifacio , mon . not .
. soc . * 419 * , 2195 ( 2012 ) .
w. aoki , h. ito and a. tajitsu , astrophys .
j. * 751 * , l6 ( 2012 ) .
w. aoki , memorie della societa astronomica italiana supplementi * 22 * , 35 ( 2012 ) .
l. monaco , p. bonifacio , l. sbordone , s. villanova and e. pancino , astron .
astrophys . * 519 * , l3 ( 2010 ) .
a. mucciarelli , m. salaris , l. lovisi , f. r. ferraro , s. lucatello and r. g. gratton , mon . not .
. soc . * 412 * , 81 ( 2011 ) .
a. frebel , w. aoki , n. christlieb , h. ando , m. asplund , p. s. barklem , t. c. beers and k. eriksson _ et al .
_ , nature * 434 * , 871 ( 2005 ) . w. aoki , a. frebel , n. christlieb , j. e. norris , t. c. beers , t. minezaki , p. s. barklem and s. honda _ et al .
_ , astrophys . j. * 639 * , 897 ( 2006 ) .
b. d. fields , ann .
nucl . part .
sci . * 61 * , 47 ( 2011 ) . c. p. deliyannis , p. demarque and s. d. kawaler , astrophys .
j. suppl .
* 73 * , 21 ( 1990 ) .
m. h. pinsonneault , t. p. walker , g. steigman and v. k. narayanan , astrophys .
j. * 527 * , 180 ( 2002 ) . m. h. pinsonneault , g. steigman , t. p. walker and v. k. narayanans , astrophys . j. * 574 * , 398 ( 2002 ) .
o. richard , g. michaud and j. richer , astrophys .
j. * 619 * , 538 ( 2005 ) .
a. j. korn _ et al .
_ , astrophys . j. * 671 * , 402 ( 2007 ) .
k. lind , f. primas , c. charbonnel , f. grundahl and m. asplund , astron .
astrophys . * 503 * , 545 ( 2009 ) .
m. kusakabe , a. b. balantekin , t. kajino and y. pehlivan , phys .
rev d * 87 * , 085045 ( 2013 ) .
d. lindley , mon . not .
. soc . * 188 * , 15p ( 1979 ) .
j. r. ellis , d. v. nanopoulos and s. sarkar , nucl .
b * 259 * , 175 ( 1985 ) .
s. dimopoulos , r. esmailzadeh , l. j. hall and g. d. starkman , astrophys .
j. * 330 * , 545 ( 1988 ) .
j. ellis , g. b. gelmini , j. l. lopez , d. v. nanopoulos and s. sarkar , nucl .
b * 373 * , 399 ( 1992 ) .
m. kawasaki and t. moroi , prog .
phys . * 93 * , 879 ( 1995 ) .
m. kawasaki and t. moroi , astrophys .
j. * 452 * 506 ( 1995 ) .
k. jedamzik , phys .
* 84 * , 3248 ( 2000 ) .
m. kawasaki , k. kohri and t. moroi , phys .
d * 63 * , 103502 ( 2001 ) .
r. h. cyburt , j. r. ellis , b. d. fields and k. a. olive , phys .
d * 67 * , 103521 ( 2003 ) .
m. kawasaki , k. kohri and t. moroi , phys .
d * 71 * , 083502 ( 2005 ) .
j. r. ellis , k. a. olive and e. vangioni , phys .
b * 619 * , 30 ( 2005 ) .
k. jedamzik , phys .
d * 74 * , 103509 ( 2006 ) .
m. kusakabe , t. kajino and g. j. mathews , phys .
d * 74 * , 023526 ( 2006 ) .
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 '' _ , tsukuba , japan , feb .
13 - 14 , 1979 , edited by o. sawada and a. sugamoto , kek report kek-79 - 18 , p. 95 .
t. yanagida , prog .
* 64 * , 1103 ( 1980 ) .
m. gell - mann , p. ramond and r. slansky , in
_ `` supergravity '' _ ( north - holland , amsterdam , 1979 ) _
eds_. d. z. freedom and p. van nieuwenhuizen , print-80 - 0576 ( cern ) .
m. fukugita and t. yanagida , phys .
b * 174 * , 45 ( 1986 ) .
t. asaka , s. eijima and h. ishida , jhep * 1104 * ( 2011 ) 011 .
t. asaka , s. eijima and a. watanabe , jhep * 1303 * ( 2013 ) 125 .
a. d. dolgov , s. h. hansen , g. raffelt and d. v. semikoz , nucl .
b * 580 * , 331 ( 2000 ) .
a. d. dolgov , s. h. hansen , g. raffelt and d. v. semikoz , nucl .
b * 590 * , 562 ( 2000 ) .
l. kawano , nasa sti / recon technical report n * 92 * , 25163 ( 1992 ) .
m. s. smith , l. h. kawano and r. a. malaney , astrophys .
j. suppl .
* 85 * , 219 ( 1993 ) .
s. sarkar , rept .
phys . *
59 * , 1493 ( 1996 ) .
r. h. cyburt _ et al .
_ , astrophys
. ser . * 189 * , 240 ( 2010 ) .
j. beringer _
[ particle data group collaboration ] , phys .
d * 86 * , 010001 ( 2012 ) and 2013 partial update for the 2014 edition ( url : http://pdg.lbl.gov ) .
a. p. serebrov and a. k. fomin , phys .
c * 82 * , 035501 ( 2010 ) .
g. j. mathews , t. kajino and t. shima , phys .
d * 71 * , 021302 ( 2005 ) .
a. serebrov , v. varlamov , a. kharitonov , a. fomin , y. .pokotilovski , p. geltenbort , j. butterworth and i. krasnoschekova _ et al .
_ , phys .
b * 605 * , 72 ( 2005 ) .
g. steigman , jcap * 0610 * , 016 ( 2006 ) .
g. audi , a. h. wapstra and c. thibault , nucl .
a * 729 * , 337 ( 2003 ) .
d. j. fixsen , astrophys . j. * 707 * , 916 ( 2009 ) .
m. kusakabe _ et al .
_ , phys .
d * 79 * , 123513 ( 2009 )
. t. shima _ et al .
_ , phys .
c * 72 * , 044004 ( 2005 ) .
t. kii , t. shima , y. nagai and t. baba , nucl .
instrum .
a * 552 * , 329 ( 2005 ) .
f. c. jones , phys .
rev . * 167 * , 1159 ( 1968 ) .
r. j. protheroe , t. stanev and v. s. berezinsky , phys .
d * 51 * , 4134 ( 1995 ) .
v. s. berezinskii , s. v. bulanov , v. a. dogiel , v. l. ginzburg and v. s. ptuskin 1990 , astrophysics of cosmic rays , ed . v. s. verezinskii and v. l. ginzburg ( new york : north - holland ) . m. kusakabe , a. b. balantekin , t. kajino and y. pehlivan , phys . lett .
b * 718 * , 704 ( 2013 ) .
m. h. reno and d. seckel , phys .
d * 37 * , 3441 ( 1988 ) .
m. kawasaki and m. kusakabe , phys .
d * 86 * , 063003 ( 2012 ) . j. m. blatt and v. f. weisskopf , _ theoretical nuclear physics _ ( dover , mineola , ny , 1991 ) .
r. h. cyburt and b. davids , phys .
c * 78 * , 064614 ( 2008 ) .
r. j. scherrer and m. s. turner , astrophys .
j. * 331 * , 19 ( 1988 ) [ astrophys .
j. * 331 * , 33 ( 1988 ) ] .
g. m. fuller , c. t. kishimoto and a. kusenko , arxiv:1110.6479 [ astro-ph.co ] . e. w. kolb and r. j. scherrer , phys .
d * 25 * , 1481 ( 1982 ) .
v. f. shvartsman , sov .
jetp lett .
* 9 * , 184 ( 1969 ) .
g. steigman , d. n. schramm and j. e. gunn , phys .
b * 66 * , 202 ( 1977 ) .
j. l. feng , a. rajaraman and f. takayama , phys .
d * 68 * , 063504 ( 2003 ) .
e. w. kolb and m. s. turner , the early universe , addison - wesley , 1990 .
m. pettini and r. cooke , mon . not .
soc . * 425 * , 2477 ( 2012 ) . t. m. bania , r. t. rood and d. s. balser , nature * 415 * , 54 ( 2002 ) .
y. i. izotov and t. x. thuan , astrophys .
j. * 710 * , l67 ( 2010 ) .
e. aver , k. a. olive and e. d. skillman , jcap * 1005 * , 003 ( 2010 ) .
k. lind , j. melendez , m. asplund , r. collet and z. magic , astron .
astrophys . * 544 * , a96 ( 2013 ) .
m. seiffert , d. j. fixsen , a. kogut _ et al .
_ , astrophys .
j. * 734 * , 6 ( 2011 ) .
d. j. fixsen , e. s. cheng , j. m. gales , j. c. mather , r. a. shafer and e. l. wright , astrophys . j. * 473 * , 576 ( 1996 ) .
n. cabibbo , phys .
* 10 * , 531 ( 1963 ) .
m. kobayashi and t. maskawa , prog .
phys . * 49 * , 652 ( 1973 ) .
d. a. dicus , e. w. kolb and v. l. teplitz , astrophys . j. * 221 * , 327 ( 1978 ) .
t. asaka , m. laine and m. shaposhnikov , jhep * 0701 * ( 2007 ) 091 .
s. hannestad , phys .
d * 70 * , 043506 ( 2004 ) .
k. ichikawa , m. kawasaki and f. takahashi , phys .
d * 72 * , 043522 ( 2005 ) .
a. kusenko , s. pascoli and d. semikoz , jhep * 0511 * ( 2005 ) 028 .
t. asaka and a. watanabe , jhep * 1207 * ( 2012 ) 112 .
l. m. johnson , d. w. mckay and t. bolton , phys .
d * 56 * , 2970 ( 1997 ) . | the primordial lithium abundance inferred from spectroscopic observations of metal - poor stars is @xmath0 times smaller than the theoretical prediction in standard big bang nucleosynthesis ( bbn ) model .
we assume a simple model composed of standard model particles and a sterile neutrino @xmath1 with mass of @xmath2 mev which decays long after bbn .
we then investigate cosmological effects of a sterile neutrino decay , and check if a sterile neutrino can reduce the primordial lithium abundance .
we formulate the injection spectrum of nonthermal photon induced by the @xmath1 decay .
we take into account the generation of electrons and positrons , @xmath3 s , and active neutrinos at the @xmath1 decay , the primary photon production via the inverse compton scattering of cosmic background radiation ( cbr ) by energetic @xmath3 , and electromagnetic cascade showers induced by the primary photons . the steady state injection spectrum
is then derived as a function of the @xmath1 mass and the photon temperature .
the @xmath1 decay produces energetic active neutrinos which are not thermalized , and @xmath3 s which are thermalized .
we then derive formulae relevant to the @xmath1 decay rates and formulae for the baryon - to - photon ratio @xmath4 and effective neutrino number @xmath5 .
the initial abundance , mass , and lifetime of @xmath1 are taken as free parameters .
we then consistently solve ( 1 ) the cosmic thermal history , ( 2 ) nonthermal nucleosynthesis induced by the nonthermal photons , ( 3 ) the @xmath4 value , and ( 4 ) the @xmath5 value .
we find that an effective @xmath6be destruction can occur only if the sterile neutrino decays at photon temperature @xmath7 kev .
amounts of energy injection at the @xmath1 decay are constrained from limits on primordial d and @xmath6li abundances , the @xmath5 value , and the cbr energy spectrum .
we find that @xmath6be is photodisintegrated and the li problem is partially solved for the lifetime @xmath8 s and the mass @xmath9 mev .
@xmath6be destruction by more than a factor of three is not possible because of an associated d over - destruction . in the parameter region ,
the @xmath4 value is decreased slightly , while the @xmath5 value is increased by a factor of @xmath10 1 . in this study ,
errors in photodisintegration cross sections of @xmath6be(@xmath11 , @xmath12)@xmath13he and @xmath6li(@xmath11 , @xmath12)@xmath13h that have propagated through literatures are corrected , and new functions are derived based on recent nuclear experiments .
it is found that the new photodisintegration rates are 2.3 to 2.5 times smaller than the old rates .
the correct cross sections thus indicate significantly smaller efficiencies of @xmath6be and @xmath6li photodisintegration .
abundances of sterile neutrino necessary for the @xmath6li reduction are much smaller than thermal freeze - out abundances .
the relic sterile neutrino , therefore , must be diluted between the freeze - out and bbn epochs by some mechanism . |
the mechanism for converting rotational energy of pulsars to electromagnetic radiation has long been the subject of active research in pulsar physics .
the recent observation of seven pulsars by the instruments aboard cgro provides even more challenge to the current pulsar theory .
these pulsars radiate high - energy photons with apparently high efficiency ( ulmer 1994 ; thompson et al . 1994 ) . one possible way to convert
the rotational energy to electromagnetic radiation is through particle acceleration by the rotation - induced electric field in the pulsar s magnetosphere .
there are several sites for gap acceleration , which include the polar gap ( ruderman & sutherland 1975 ; arons & scharlemann 1979 ; michel 1974 ; fawley , arons & scharlemann 1977 ) , the slot gap ( arons 1983 ) , and the outer gap ( cheng , ho & ruderman 1986 ) . in the polar gap model , the acceleration region is located near the polar cap while in the outer gap model , the acceleration occurs near the light cylinder . in the slot gap model
, the acceleration is at the boundary between the open and closed field lines .
there is observational evidence that polar caps of young pulsars can be hot with effective temperature @xmath3 ( gelman 1991 ; greiveldinger et al .
although the exact value of @xmath4 is rather uncertain , and may depend on the model of the atmosphere above the polar cap ( romani 1987 ) , the hot polar cap appears to be a plausible consequence of polar cap heating as the result of particle acceleration ( cheng & ruderman 1977 ; arons & scharlemann 1979 ; luo 1996 ) .
one of the important consequences of hot polar caps is that it provides an alternative mechanism for controlling the gap , i.e. mainly through pair cascades initiated by compton scattered photons ( sturner 1995 ; luo 1996 ) .
inverse compton scattering in the polar cap region is strongly modified by the magnetic field in that the scattering cross section is enhanced by the cyclotron resonance ( herold 1979 ) , i.e. in the electron rest frame , soft photons have energies close to the cyclotron energy @xmath5 in @xmath6 with @xmath7 the critical field .
this process is referred to as resonant inverse compton scattering ( rics ) . because of the resonance , the effective cross section is greatly enhanced in comparison with the ordinary compton scattering . for gap acceleration , we can define its efficiency as the ratio of the total voltage across the gap to the maximum voltage across the polar cap ( i.e. for an empty magnetosphere ) .
the efficiency of the gap acceleration is controlled by the pair production and thus strongly depends on the effective temperature of the polar caps . in this paper
, we consider the constraint imposed by rics on the efficiency of polar gap acceleration .
we assume that free emission of charges from polar caps is allowed , and consider both the polar region in which the outflowing charges are electrons ( @xmath8 ) and that in which the outflowing charges are heavy ions or positrons ( @xmath9 ) . in the polar gap at which @xmath9 , since the energy loss of ions due to rics is negligibly small , they can be accelerated by the full potential drop across the gap .
we derive the condition for the ions to carry most of the particle luminosity .
accelerated ions may produce pairs in the anisotropic thermal photon field .
we calculate numerically the distance from the polar cap at which the ions start to produce pairs and compare it to the gap length constrained by pair production by rics .
for a hot polar cap which emits soft photons , rics can constrain the maximum energies of electrons ( or positrons ) accelerated in the polar gap in the following two ways : the energy loss competes with acceleration , or by limiting the acceleration length through pair production . for a given accelerating potential , @xmath10 , the lorentz factor of an electron ( or positron ) at a distance , @xmath11 , to the polar cap
is given by ( dermer 1990 ; luo 1996 ) @xmath12,\eqno(1)\ ] ] where @xmath13 is the initial lorentz factor , @xmath14 is the compton wavelength , @xmath15 is the effective temperature ( in @xmath6 ) , @xmath16 is the star s radius , @xmath17 is the thomson cross section , @xmath18 is the fine structure constant , @xmath19 $ ] , and @xmath20 is the maximum angle between a photon and the particle motion . for thermal emission from the entire neutron star s surface
we have @xmath21^{1/2}/(x+1)\}$ ] , which corresponds to propagation of photons at directions tangential to the surface . for a hot polar cap ,
this is valid only for @xmath22 ( where @xmath23 and @xmath24 is the pulsar period ) ; for @xmath25 , the angle is given by @xmath26 $ ] , corresponding to photons from the polar cap rim on which the field lines have colatitude angle @xmath27 relative to the magnetic pole .
the energy loss due to rics is described by the third term on the right hand side of equation ( 1 ) .
the maximum lorentz factor corresponds to the value at @xmath28 which is the gap length determined by the pair production , e.g. due to rics or due to thermal photons interacting with heavy ions , whichever comes first .
equation ( 1 ) can be solved numerically with the boundary condition @xmath29 .
the model for the accelerating potential strongly depends on the physical conditions on the polar cap .
if polar caps are sufficiently hot , we may have free emission of charges .
the induced electric field causes charge flow , which tends to screen out the field .
however , the charge screening is always insufficient , e.g. due to field line curvature ( arons & scharlemann 1979 ) , and hence charge depletion occurs ( the charge density deviates from goldreich - julian ( g - j ) density ) . an electric potential exists in the charge - depletion region ( often called the gap ) . as an example , we consider the potential ( e.g. arons & scharlemann 1979 ; arons 1983 ) @xmath30\,g(x)\,\sin\psi\,\sin i,\eqno(2)\ ] ] where @xmath31 is the angle of the field lines relative to the magnetic pole , @xmath32 is the azimuthal angle of the open field line flux tube , @xmath33 is the distance to the polar cap ( in @xmath16 ) , @xmath34 is the angle between the rotation axis and the magnetic pole , and @xmath35 is given by @xmath36 , @xmath37\ \ { \rm for}\ \ x\gapprox0.5\theta_c$ ] . according to arons ( 1983 ) , the angle
@xmath32 is in the ranges , @xmath38 for @xmath39 ( if the outflowing charges are ions ) , and @xmath40 for @xmath41 ( if the outflowing charges are electrons ) .
the potential given by ( 2 ) is due to charge density decrease along field lines ( i.e. it becomes less than the g - j density ) because of field line curvature .
plots of the lorentz factor as a function of distance from the polar cap are shown in figure 1 .
we assume @xmath42 at @xmath43 . as shown in the figure
, energy loss due to rics can compete with acceleration only for high polar cap temperature @xmath44 . for @xmath45
, the particle energy is constrained mainly through pair production by rics .
there are other models of the space - charge - limited potential such as the model discussed by michel ( 1974 ) , also by fawley , arons & scharlemann ( 1977 ) and cheng & ruderman ( 1977 ) , which do not rely on field line curvature and in general gives a lower voltage .
the potential discussed by michel ( 1974 ) arises from the deviation of the charge density from the g - j value because of particle inertia , i.e. the velocity @xmath46 and the magnetic field @xmath47 have different scaling with @xmath33 such that the charge density can not be maintained at the g - j density all the way along open field lines .
the corresponding potential can be described by ( fawley , arons & scharlemann 1977 ) @xmath48 with @xmath49 , @xmath50 and @xmath51 . for @xmath52 ,
the electric field drops off faster than @xmath53 .
the plot of the lorentz factor as a function of @xmath11 is shown in figure 2 for @xmath54 and @xmath55 .
the electric potential for @xmath52 is assumed to be constant , @xmath56 . for @xmath55 ,
the energy loss is so severe that it continues beyond the distance @xmath52 , resulting in a sudden decrease in @xmath57 at @xmath58 . as in figure 1 , for moderately hot polar caps ( @xmath59 ) , the particle energy is constrained by electron - positron cascades started by rics . the main mechanism for a high energy photon ( emitted by an electron or positron either through curvature or inverse compton scattering ) to produce a pair is single photon decay in a strong magnetic field .
to produce a pair , the photon must satisfy the threshold condition @xmath60 where @xmath61 is the photon energy ( in @xmath6 ) , @xmath62 is the angle between the photon propagation and the magnetic field . when this condition is satisfied , the number of pairs produced depends on the opacity , @xmath63 , which is an integration of the absorption coefficient ( for a photon being converted into a pair ) along the distance that photon travels .
we assume that a photon produces at least one pair if @xmath64 . with this condition
, we can write down the threshold condition for producing at least one pair through rics ( luo 1996 ) as @xmath65 the right - hand side of ( 4 ) strongly depends on @xmath66 . if the above condition is satisfied , pair cascade can be initiated by rics .
the threshold condition is represented by the dotted curves in figures 1 and 2 .
note that for each curve of @xmath67 vs. @xmath68 , the location of an intersect ( with the dotted curves ) point defines a gap length , represented by @xmath69 .
( if the curve has two intersecting points , the one that is the nearest to the polar cap is relevant . )
electron positron pair cascades can be started by accelerated ions in thermal photon fields .
a photon can decay into an electron positron pair in the coloumb field of a nucleus when the center of momentum frame energy exceeds the rest mass of the nucleus plus two electrons , that is @xmath70 where @xmath71 is the mass of the nucleus .
this condition can be satisfied for soft photons from the polar cap in the presence of accelerated ions .
we can estimate the mean free path of an ion before producing an average one pair .
if we believe one pair per primary particle is enough to short out the electric field , this mean free path length , represented by @xmath72 , can be identified as the gap length ( figures 3 and 4 ) .
numerical evaluation of the mean free path for an accelerated ion traversing an isotropic thermal photon field was given by protheroe ( 1984 ) , and protheroe & johnson ( 1996 ) .
the injection rate of pairs for a given energy of iron nuclei was numerically calculated by bednarek & karakula ( 1995 ) for an anisotropic photon field from the polar cap .
we calculate number of pairs produced by iron nuclei accelerated along open field lines by the potentials given by eqs .
( 2 ) and ( 3 ) , and the mean free path corresponds to the average distance at which one pair is produced per ion nucleus . the mean cumulative number of pairs produced in the radiation field of the polar cap ( or whole neutron star ) at temperature @xmath4 by a nucleus of energy @xmath73 traveling along the neutron star magnetic axis from the pole to distance @xmath11 above the polar cap surface is given by , @xmath74}^{\cos[\theta_{\rm min}(e,\varepsilon ) ] } ( 1 - \beta \cos \theta ) \sigma(s ) \ , d\cos\!\theta \ , d\varepsilon \ , dx , \eqno(6)\ ] ] where @xmath75 is the differential photon number density of photons of energy @xmath61 in black body radiation at temperature @xmath4 , and @xmath76 is the total cross section for pair production ( protheroe 1997 ) by a the nucleus at center of momentum ( cm ) frame energy squared , @xmath77 , given by @xmath78 where @xmath31 is the angle between the directions of the nucleus and photon .
note that @xmath79 is determined by the threshold condition ( 5 ) .
plots of accumulative number of pairs produced per ion through interaction with thermal photons as a function of distance from the polar cap are shown in figure 4 . for @xmath80 and @xmath81 .
compared to rics , in this case we have @xmath82 .
although we use a high @xmath66 , the process strongly depends on @xmath24 and @xmath4 .
the efficiency of the polar gap is described by the ratio of the potential drop across the gap , @xmath83 , to the maximum potential across the polar cap with an empty magnetosphere , @xmath84 , viz .
@xmath85 where @xmath86 , @xmath87 , and @xmath88 is the length of the polar gap in units of @xmath16 .
let @xmath89 and @xmath90 be the accelerated particle and spin - down luminosities , respectively . if the energy loss of accelerated particles in the gap is not important , we have @xmath91 since we can write @xmath92 and @xmath93 where @xmath94 is the injection rate calculated from the g - j density @xmath95 . note that the efficiency defined by ( 8) is similar to that used by arons ( 1996 ) . for the presently available models for accelerating potentials
, we always have @xmath96 . using ( 8)
, we may estimate the efficiency for a given model potential .
for the potential described by eq .
( 2 ) and assuming @xmath97 , we have the maximum efficiency @xmath98 thus , even when we use @xmath99 , the maximum efficiency is @xmath100 .
when a pair cascade occurs , we usually have @xmath101 ( for the polar gap ) and hence @xmath102 .
one may estimate the efficiency for the potential given by eq .
( 3 ) from @xmath103 with @xmath104 .
although we have a higher @xmath105 for the arons & scharlemann model , the acceleration by the potential given by eq .
( 3 ) is the more effective for @xmath106 .
the evaluation of @xmath88 depends on the specific mechanism for initiating a pair cascade and whether the pair plasma produced through the cascade is dense enough to short out part of the electric field .
we consider three mechanisms for starting a pair cascade : rics , curvature radiation , and inelastic scattering of ions by thermal photons from the polar cap .
the corresponding lengths are @xmath69 , @xmath107 and @xmath72 . in general
, we have @xmath108 ( e.g. luo 1996 ) . for moderately hot polar caps with effective temperature @xmath109 ,
the energy loss due to rics is not important compared to acceleration but the photons produced through rics can start a pair cascade at the distance less than @xmath108 , @xmath72 .
thus , the effect of rics is to reduce the gap acceleration efficiency . for space - charge - limited flow
, the composition of outflowing charges at the poles with @xmath110 and @xmath111 can be different . for @xmath110 ,
the primary particles consist mainly of electrons , and for @xmath111 , the main components are heavy ions or positrons . in the ion zone ,
the gap length is constrained by the pair production by positrons through rics .
the possible source of positrons was discussed by cheng & ruderman ( 1977 ) .
for moderately hot polar caps , the gaps at both types of pole have a similar efficiency . for sufficiently hot polar caps and a superstrong magnetic field ( @xmath112 ) , the energy loss of electrons or positrons due to rics can be important , and may prevent them starting a pair cascade in the region close to the polar cap . since the cross section is @xmath113 for iron nuclei ( @xmath114 ) ,
the energy loss due to rics for ions is negligibly small , and they can be continuously accelerated until distance @xmath115 .
thus , the acceleration efficiency increases significantly . as an example , for the potential ( 2 ) with @xmath116 and @xmath117 , the gap length is about @xmath118 and the maximum energy of electrons ( or positrons ) is @xmath119 .
this gives the particle luminosity @xmath120 where @xmath94 is the injection rate of primary electrons ( or positrons ) .
for hot polar caps with @xmath121 , from figures 1 , we obtain @xmath122 . from figure 4 ,
pair production by ions interacting with thermal photons is important only for the case where the whole star s surface has @xmath123 ( the thick solid curve ) , and this occurs at the distance @xmath124 .
thus , the gap length is controlled by rics . from figure 1
, we have @xmath125 for positrons
( or electrons ) and @xmath126 for ions ; positrons ( or electrons ) and ions can be accelerated to the maximum energy @xmath127 and @xmath128 , respectively .
hence , we have much a higher particle luminosity @xmath129 and most of @xmath89 are carried by ions .
from observations , several pulsars may have polar caps with effective temperature @xmath130 though the exact value of the temperatures is rather uncertain ( for a review , see e.g. , gelman 1991 ) . for moderately hot polar caps and typical b fields ( @xmath131 ) ,
the energy loss rate is not rapid enough to compete with the acceleration , and rics controls the polar gap only through pair production .
this in general reduces the gap length , and hence reduces the acceleration efficiency by at least one order magnitude , imposing a severe constraint on the energetics of the polar gap . for young pulsars with superstrong magnetic fields
@xmath132 , the gap acceleration efficiency can increase provided that the polar cap ( @xmath111 ) is hot ( @xmath133 ) and for electrons ( or positrons ) , the energy loss due to rics can dominate over the acceleration .
one of cgro pulsars , psr 1509 - 58 , has magnetic field @xmath134 ( the magnetic field at the pole can be twice this value ) .
since neutron stars are hot at birth ( after supernova explosion ) , their polar caps can be hot as result of self - sustained polar cap heating at very young age ( e.g. luo 1996 ) , e.g within hundred years or so after birth .
these high field pulsars , assuming they were born with high magnetic fields , must have undergone a phase when thermal emission had an important effect on pulsar electrodynamics .
we suggest that for high field pulsars there can be four phases during the their lifetime as discussed below .
( i ) at very young age ( within a few years after supernova explosion ) when the surface or polar cap temperature is high , energy loss due to rics is so severe that a pair cascade due to electrons or positrons can not occur in the region close to the polar cap . in the ion zone ( @xmath111 )
, accelerated ions may carry most of the particle luminosity .
for heavy ions , photodisintegration may occur in the thermal photon field from the polar cap . at this stage , since pulsars also have a very large spin - down luminosity , they are a potential source of high energy hadrons and electromagnetic radiation .
( ii ) as pulsars cool down , the energy loss rate due to rics becomes less than acceleration rate .
however , rics may still be the dominant mechanism for pair production to constrain the gap length ( luo 1996 )
. the high field pulsar , psr 1509 - 58 , may currently be at this stage since the polar cap temperature inferred from the observation is about @xmath135 ( e.g. kawai et al . 1991 ) .
its gap length is shorter than that controlled by pair production by either curvature radiation or relativistic ions in the photon field .
thus , the acceleration efficiency is less than in phase ( i ) .
( iii ) as the surface temperature or polar cap temperature becomes much lower , pair production by rics is not important .
the main mechanism for initiating pair cascades is curvature radiation . as the period increases , the gap length increases and hence the efficiency also increases .
however , at this stage , the pulsar spin - down luminosity is lower than at their early young age .
( iv ) the period reaches the critical period at which pair production due to curvature radiation is no longer operative .
although the gap acceleration efficiency is now high ( no pairs to screen the electric field ) , the luminosity is very low . in summary , for very young pulsars with superstrong magnetic fields ( @xmath136 ) and hot polar caps ( @xmath137 ) , the severe energy loss due to resonant inverse compton scattering can prevent electrons or positrons starting a pair cascade in the region close to the polar cap but ions can be accelerated by the full potential drop across the gap ( provided that free emission of ions can occur ) .
we then have the ions carrying most of the particle luminosity .
thus , very young pulsars with superstrong @xmath66 field can be a potential source of high - energy hadrons and gamma - rays .
in contrast , for pulsars with moderately hot polar caps ( @xmath3 ) , the effect of resonant inverse compton scattering is to reduce the gap acceleration efficiency .
pair production by accelerated ions interacting with thermal photons strongly depends on @xmath24 and @xmath4 , and in general we have @xmath82 and hence the polar gap length is controlled by rics .
ql thanks australian research council ( arc ) for financial support through a fellowship .
we thank wlodek bednarek for help discussion .
arons , j. 1996 , a&as , 120 , 49 .
arons , j. 1983 , apj , 266 , 215 .
arons , j. & scharlemann , e. t. 1979 , apj , 231 , 854 .
bednarek , w. & karakula , s. 1995 , in proc .
. ray conf .
( rome ) , v.2 , p. 279 .
cheng , a. f. & ruderman , m. a. 1977 , apj , 214 , 598 .
cheng , k. s. , ho , c. & ruderman , m. a. 1986 , apj , 300 , 500 .
dermer , c. d. 1990 , apj , 360 , 197 .
fawley , w. m. , arons , j. & scharlemann , e. t. 1977 , apj , 217 , 227 .
greiveldinger , c. et al .
1996 , apj , 465 , l35 .
herold , h. 1979 , phys .
d19 , 2868 .
kawai , n. et al .
1991 , apj , 383 , l65 .
luo , q. 1996 , apj , 468 , 338 .
michel , f. c. 1974 , apj , 192 , 713 .
gelman , h. 1991 , in thermal emission of pulsars , in neutron stars : theory and observation , ed
. j. ventura & d. pines ( dordrecht : kluwer ) , 87 .
protheroe , r. j. 1984 , nature , 310 , 296 .
protheroe , r. j. 1997 , in towards the millennium in astrophysics : problems and prospects , eds .
shapiro and j.p .
wefel ( world scientific , singapore ) , in press .
protheroe , r. j. & johnson , p. a. 1996 , astroparticle phys . , 4 , 253 .
romani , r. w. 1987 , apj , 313 , 718 .
ruderman , m. & sutherland , p. g. 1975 , apj , 196 , 51 .
sturner , s. j. 1995 , apj , 446 , 292 .
thompson , d. j. et al .
1994 , apj , 436 , 229 .
ulmer , m. p. 1994 , apjss , 90 , 789 . | it is shown that for moderately hot polar caps ( with effective temperature of @xmath0 k ) , the efficiency of polar gap acceleration is lower compared to the case in which the polar caps are relatively cool and inverse compton scattering plays no role in controlling the gap . for young pulsars with superstrong magnetic fields ( @xmath1 ) and hot polar caps ( with temperature of @xmath2 ) , because of the energy loss of electrons or positrons due to resonant inverse compton scattering in the vicinity of polar caps , pair cascades occur at distances further away from the polar cap , and in this case , we have a relatively high acceleration efficiency with ions carrying most of the particle luminosity .
= 2em = 17.5 cm = 24.6 cm = -2.5 cm = -1.0 cm = -1.0 cm department of physics and mathematical physics , + the university of adelaide , sa 5005 + * keywords : * acceleration of particles pulsars : general star : neutron scattering |
euclidean feynman path integral uniquely identifies the ground state of a field theory ( vaccum)@xcite .
the feynman functional integral is defined as the limit of ordinary integrals defined on discrete set of points in a four dimensional box , when the number of points is sent to infinity filling the box densely ; the size of the box is then sent to infinity to cover the whole space time .
lattice formulation is an approssimant in this sequence .
qcd has an uv fixed point at @xmath0 ( asymptotic freedom ) : as @xmath1 the physical length scale @xmath2 increases in units of the lattice spacing @xmath3 @xmath4 the density of lattice points in physical units goes large . on the other hand a mass gap exists in the theory , which makes the infinite volume limit well defined as a thermodynamical limit .
lattice is a good approximant of qcd if @xmath5 with @xmath6 the size of the lattice .
the above argument is a strong indication that qcd most probably exists as a field theory in the constructive sense , and can be defined as a limit of lattice formulation as @xmath7 , and @xmath6 is such that the inequality ( [ eq2 ] ) is satisfied .
the textbook quantization of qcd is based on perturbation theory , and the ground state is fock vacuum of quarks and gluons .
phenomenological evidence exists@xcite that the qcd vacuum is not the perturbative vacuum .
the instability of fock s vacuum is possibly the origin of the non borel summability of the perturbative expansion@xcite .
perturbation theory apparently works at small distances but is not well defined and is unable to describe large distance physics .
the most clear evidence that fock vacuum is not a good approximation to qcd ground state , is that its elementary excitations , quarks and gluons , have never been observed as free particles .
this phenomenon is known as confinement of colour and is one of the most intriguing properties of qcd .
the study of the mechanism of confinement is an important chapter of the investigation of the structure of qcd vacuum on the lattice , and will be the object of this talk . in sect.2 i will review the experimental evidence for confinement . in sect.3
i will discuss the phenomenology of the deconfinement transition as observed in numerical simulations on the lattice .
this will naturally lead us to the idea of duality , which will be the object of sect.4 .
progress in understanding confinement as dual superconductivity of the vacuum will be reviewed in sect.5
confinement is defined as the absence of colored particles in asymptotic states .
the existing experimental evidence for confinement is based on the negative result of searches of fractionally charged particles ( quarks ) in particle reactions and in nature .
the cross section @xmath8 for the inclusive production of quark or antiquark in the process @xmath9 at c.m .
energies @xmath10 gev has an upper limit@xcite @xmath11 to be compared with the total cross section @xmath12 at the same energy @xmath13 .
the ratio @xmath14 has then the upper bound @xmath15 while the expected value in the absence of confinement is a sizable fraction of unity .
the negative result of the search of fractionally charged particles in ordinary matter by millikan - like experiments gives an upper limit for the ratio @xmath16 of the quarks abundance @xmath17 to nucleons abundance @xmath18 @xmath19 corresponding to the analysis of @xmath20 gr of matter . in the absence of confinement
the expectation for @xmath16 , @xmath21 , in the standard cosmological model is @xmath22@xcite .
again @xmath23 a ratio smaller than @xmath24 , if different from zero , would be too small to have a natural explanation in any theory .
the most natural interpretation is then that those ratios are strictly zero , or that confinement is an absolute property of the vacuum based on a symmetry@xcite .
the question is : what symmetry of qcd vacuum prevents quarks to exist as free particles ? as for gluons , they have no such characteristic signature as a fractional charge , and their identification is not clearly feasible .
no experimental data exist on gluon confinement .
we shall define anyhow confinement as absence of any colored particle as a free particle .
qcd at finite temperature can be studied on the lattice .
the partition function @xmath25 of a field with action @xmath26 $ ] is equal to the feynman integral in euclidean space , with the time direction extending from 0 to @xmath27 and periodic boundary conditions in time for bosons , antiperiodic for fermions @xmath28 e^ { -\beta\!\int \ ! d^3x\int_0^{1/t}\hskip-5pt
dx_0 s_e[\phi(\vec x , x_0 ) ] } \ ] ] in lattice qcd this corresponds to having a lattice of size @xmath29 , with @xmath30 , and the temperature is given by the inverse of the temporal extension @xmath31 ( @xmath3 the lattice spacing ) .
the value of @xmath3 in physical units depends on the coupling constant ( @xmath32 ) via renormalization group @xmath33 with @xmath34 hence @xmath35 as a consequence of asymptotic freedom low temperature ( confinement ) corresponds to strong coupling ( large @xmath36 ) or to disorder in the language of statistical mechanics , high temperature corresponds to order or to weak couling .
if confinement is due to a symmetry , it has to be a symmetry of the disordered phase .
this naturally leads to the idea of duality , which will be the object of sect.3 .
the deconfinement transition is detected on the lattice in pure gauge theory , by looking at the correlator @xmath37 where @xmath38 , the polyakov line , is the trace of the parallel transport along the time axis across the lattice and back via periodic boundary conditions @xmath39 the static potential between a @xmath40 pair is given by @xmath41 by cluster property , at large distances @xmath42 a critical temperature @xmath43 is observed such that @xmath44 confinement is related to the presence of a linear potential at large distances : @xmath45 is known as string tension . for @xmath46 gauge theory the phase transition at @xmath43 is second order , and belongs to the universality class of the @xmath47 ising model . @xmath48@xcite . for @xmath49 gauge theory
the transition is weak first order , and @xmath50@xcite . in the presence of quarks the symmetry @xmath51 of which
@xmath6 is an order parameter , is not a symmetry any more . for massless
quarks a chiral symmetry exists above some temperature @xmath43 , which is spontaneously broken for @xmath52 , the pseudoscalar octet being the goldstone particles . for @xmath53
the chiral symmetry is again explicitely broken .
it is qualitatively clear that confinement can produce a breaking of chiral symmetry : in a bag model chirality is inverted in the reflection on the confining wall .
numerical indications also exist that the two transitions take place at the same temperature .
however the overall situation is not satisfactory .
first of all if a phase exist in which color is confined , and a phase at higher temperature in which quarks and gluons are free particles , an exact order parameter should exist for this transition .
in addition a strong theoretical hint exists that , when the number of colors @xmath54 is sent large , with the constraint @xmath55 fixed , a limiting theory is defined , which does not differ much in its physical content from the realistic theory where @xmath56 .
the expansion in @xmath57 should be a convergent expansion@xcite . in this philosophy quark loops
are non leading @xmath58 and therefore the physics of confinement , i.e. the symmetry , should be the same as in quenched approximation .
duality is a deep concept in field theory , in statistical mechanics , in string theory .
it applies to systems in @xmath59 dimensions having non local topological excitations in @xmath60 dimensions .
two complementary descriptions can be given of such systems . a _ direct _ description , in terms of local fields @xmath61 , in which topological excitations @xmath62 are non local .
symmetry is described by order parameters , the v.e.v .
@xmath63 of the fields @xmath61 .
this description is convenient in the ordered phase , or weak coupling regime .
a _ dual _ description in which @xmath62 are local operators , and @xmath61 non local excitations .
symmetry is described by disorder parameters @xmath64 .
the dual coupling is @xmath65 .
this description maps the strong coupling regime of the direct theory into the weak coupling regime of the dual .
therefore it is convenient in the strong coupling regime .
the prototype system@xcite with duality is the @xmath66 dimensional ising model : there the field @xmath61 is the spin variable @xmath67 .
the dual configurations @xmath62 are 1 dimensional kinks .
the dual description is again an ising model in which the creation operator of a kink is @xmath68 and the dual boltzman factor @xmath69 is defined by the relation @xmath70 or @xmath71 . in the model @xmath72
plays the role of the coupling constant .
other examples are * the 3d @xmath73 model@xcite , which belongs to the class of universality of liquid @xmath74 .
the topological excitations are abelian vortices . * the heisenberg magnet , where the topological excitations are @xmath75 weiss - domains@xcite .
* the @xmath76 susy qcd@xcite where the excitations are monopoles .
* superstring @xmath77 theories@xcite . * compact @xmath78 gauge theories , where the topological excitations are monopoles@xcite .
the question is : what are the dual excitations in qcd ?
two proposals exist in the literature , both due to g. thooft@xcite . * _ monopoles_. monopole condensation in the vacuum produces dual superconductivity , and confinement of electric charges via abrikosov flux lines ( meissner effect ) .
monopoles are defined by a procedure named `` abelian projection''@xcite based on the choice of a local operator in the adjoint representation , as discussed below . in a sense
the mechanism is largely undefined , since there is a continuous infinity of choices for the operator used to define the monopoles and their interrelation is not understood .
a guess is that all these choices are physically equivalent@xcite .
+ the positive feature is that confinement is explained in a simple form : the chromoelectric field of a @xmath79 pair is channeled into an abrikosov flux tube with energy proportional to the length , i.e. to the distance @xmath80 between the pair@xcite @xmath81 @xmath45 , the energy per unit length of the tube , is the string tension . * _ vortices_. a vortex is a magnetic defect associated to a closed line @xmath82 .
the operator @xmath83 which creates a vortex at some time @xmath84 , obeys the following algebra , with the operator @xmath85 creating a wilson loop along the line @xmath86 @xmath87 where @xmath88 is the winding number of the lines @xmath89 .
it can be shown that , whenever @xmath90 obeys the area law , @xmath91 , as the loop goes large , @xmath92 obeys the perimeter law , and viceversa if @xmath92 obeys the area law , @xmath90 obeys the perimeter law .
+ area law implies that @xmath93 , the expectation value of the polyakov line vanishes .
one can define a dual polyakov line @xmath94 wich is a vortex along a straight path crossing space from @xmath95 to @xmath96 . then for @xmath52 , @xmath97 , for @xmath98 , @xmath99 .
@xmath100 is a disorder parameter for confinement@xcite .
the definition of monopoles is associated to the choice of an operator @xmath101 in the adjoint representation : we will speak of @xmath46 gauge group for simplicity , but the procedure is easily generalized to @xmath102 .
let @xmath103 be the direction of @xmath104 in color space .
a gauge invariant field strength @xmath105 can be defined as@xcite @xmath106 @xmath107 @xmath108 the two terms in the definition ( [ eq11 ] ) are separately gauge invariant and color singlets : they are arranged in such a way that bilinear terms @xmath109 or @xmath110 cancel .
indeed @xmath111 the gauge transformation bringing @xmath61 in the same direction ( say ( 0,0,1 ) ) in color space is called abelian projection .
after abelian projection @xmath112 the source of the dual tensor @xmath113 is a magnetic current @xmath114 and is conserved . the corresponding @xmath78 magnetic symmetry can either be wigner , and then a magnetic charge operator is defined and hilbert space is superselected , or be higgs broken , which implies the existence of at least one magnetically charged operator @xmath62 such that @xmath115 .
@xmath115 implies dual superconductivity .
notice that in a noncompact formulation @xmath116 ( bianchi identities ) . in principle the existence of dual superconductivity can be investigated in any abelian projection by looking at the v.e.v .
@xmath64 of an operator creating a magnetic charge in that abelian projection .
once the dual topological excitations are identified the disorder parameter can be constructed as the v.e.v . of their creation operator @xmath62 .
the construction of @xmath117 $ ] interms of the field of the direct description @xmath61 is an explicit realization of the duality transformation .
the guiding idea goes back to ref.@xcite and amounts to a traslation of the fields in the schrdinger picture by the classical topological configuration . in the same way as @xmath118 @xmath119 with @xmath120 the conjugate momentum to the field @xmath61
@xmath121 = i \delta^3(\vec x - \vec y),\ ] ] adds @xmath122 to the field configuration latexmath:[\[\mu(\vec x)|\phi(\vec x)\rangle = construction to a compact formulation of the theory as in qcd on the lattice is far from trivial , but has been done@xcite . a correct , gauge invariant , definition of the operator @xmath62 for abelian projected monopoles exists@xcite .
the resulting operator is a dirac like magnetically charged and gauge invariant operator , which can then have non vanishing v.e.v . without violating gauge invariance@xcite .
the construction has been checked on systems already studied and understood by other methods , starting with the 2d ising model@xcite , where our operator @xmath62 which creates kinks does indeed coincide with the dual variable @xmath124 of ref.@xcite . for the 3d @xmath73 model
the phase transition was always conjectured to be produced by condensation of vortices , but that condensation was detected in numerical simulations by observing an increase of the number of vortices in condensed phase : our construction shows that @xmath125 , or that the number of vortices is not defined in the vacuum of the condensed phase@xcite .
for the heisenberg model we have found@xcite that the curie transition can be looked as an order disorder transition , and that the high temperature phase is ordered in the dual language , by condensation of non abelian @xmath126 vortices .
a numerical problem has also been solved ; @xmath62 defined by eq.([eqmu ] ) is the exponential of an integral over the spatial volume : therefore it fluctuates typically as @xmath127 , which is a wild fluctuation .
in fact the disorder parameter @xmath64 is the ratio of two partition functions @xmath128 and has the fluctuations typical of a partition function .
@xmath129 is the ordinary partition function of the theory , and @xmath130 corrisponds to a modification of the action coming from the exponent in @xmath62 , eq.([eqmu ] ) . a way out of this difficulty consists@xcite in considering instead of @xmath64 the quantity @xmath131 @xmath132 , which contains all the relevant information of @xmath64 .
a typical shape of @xmath64 is shown in fig.1 .
the corresponding @xmath133 is shown in fig.2 .
the peak seats on the location of the phase transition . for compact @xmath78 .
] for compact @xmath78 ] the phase transition in principle takes place at infinite volume .
the limit can be performed by a finite size scaling analysis as follows ; if @xmath135 is the critical index of @xmath62 and @xmath136 the reduced temperature , @xmath137 then @xmath138 the functional dipendence is dictated by the fact that @xmath61 is dimensionless : @xmath3 is the lattice spacing , @xmath2 the correlation length , @xmath6 the extension of the lattice . as @xmath139 , @xmath2 goes large with a critical index @xmath140 @xmath141 @xmath142 can be approximated with 0 ( scaling limit ) , and @xmath143 can be traded with @xmath144 . hence @xmath145 and @xmath146 @xmath147 is a universal function of @xmath144 , i.e. it scales when plotted versus @xmath144 .
since this scaling law is valid for the appropriate value of @xmath140 and @xmath148 , the determination of @xmath140 , @xmath148 and @xmath135 is then possible .
@xmath64 , or better @xmath133 , has been studied for pure gauge @xmath46 and @xmath49 gauge theories across the deconfining phase transition , in a number of different abelian projections .
* @xmath115 in the confined phase ( @xmath149 ) . * @xmath150 , @xmath151 for @xmath152 , @xmath153 being the spatial extension of the lattice .
this means @xmath154 at @xmath98 in the thermodynamical limit . *
@xmath155 * @xmath43 and @xmath140 can be determined , together with @xmath135 from the finite size scaling analysis sketched in sect.3 .
@xmath43 agrees with the determination made using the traditional order parameter @xmath93 , and so does @xmath140 . for @xmath46 @xmath157 in agreement with the expectation that the transition belongs to the universality class of the 3d ising model , @xmath158 .
the independence of the result from the choice of the abelian projection has been checked by performing the analysis in a number of different abelian projections , but also averaging on a very large number ( infinite ) of abelian projections@xcite .
the same disorder parameter @xmath64 can be used in full @xmath161 ( including dynamical quarks ) where it is well defined , contrary to @xmath162 whose symmetry @xmath51 is explicitely broken by the very presence of quarks and to @xmath163 , whose symmetry is broken by quark masses .
the analysis for full qcd is in progress and of course requires a big computational effort .
there the dependence on the masses of the quarks makes the finite size scaling analysis , eq.([eqrho2 ] ) more complicated .
preliminary results@xcite , however , indicate that also in full qcd there is a transition from dual superconductor to normal across the transition .
if these preliminary indications get confirmed by the quantitative analysis which is on the way , we would have a good order parameter for confinement , even in the presence of ( massive ) quarks . in principle
the two transitions , chiral and deconfinement , could take place at different temperatures ; however all the existing indications are that they coincide .
moreover this would also reconcile confinement with the @xmath164 limit . in that respect also
an analysis of quenched gauge theories at @xmath165 is on the way , to perform the brute force check of the limit .
however our results are still far from complete .
they indicate that , whatever the dual fundamental excitations are , they carry magnetic charge in all the abelian projections . a real theoretical breakthrough would be to identify such excitations , and to write an effective lagrangian for them to describe qcd in the confined phase . 99 r.p .
feynman , rev .
( 1948 ) 367 .
shifman , a.i .
veinshtein , v.i .
zakharov , nucl .
b147 ( 1979 ) 385,448,519 .
mller , nucl .
b250 ( 1985 ) 327 .
review of particle physics , e.p.j .
15 ( 2000 ) .
l. okun , leptons and quarks , norh holland ( 1982 ) .
g. thooft , nucl .
phys . b138 ( 1978 ) 1 .
j. engels , f. karsch , h. satz , i. montway , nucl .
b205 ( 1982 ) 239 .
b. beinlick , f. karsch , e. laerman , a. peikart , eur .
j. c6 ( 1999 ) 133 .
g. thooft , nucl .
b72 ( 1974 ) 461 .
kramers , g.h .
wannier , phys . rev . 60 ( 1941 ) 252 .
l. kadanoff , h. ceva , phys .
rev . b3 ( 1971 )
3918 . g. di cecio , a. di giacomo , g.paffuti , m. trigiante , nucl .
phys . b489 ( 1997 ) 739 .
a. di giacomo , d. martelli , g. paffuti , phys .
d ( 1999 ) 094511 .
n. seiberg , e. witten , nucl .
b431 ( 1994 ) 484 .
t. banks , w. fischler , s.h .
shenker , l. susskind : phys .
d55 , ( 1997 ) , 5112 .
j. frhlich , p.a .
marchetti , comm .
( 1987 ) 343 .
a. di giacomo , g. paffuti , phys .
d56 ( 1997 ) 6816 .
g. thooft , nucl . phys .
b190 ( 1981 ) 455 .
g. thooft , high energy physics , eps international conference , palermo ( 1975 ) , eds a. zichichi .
s. mandelstam , phys .
23c ( 1976 ) 245 . l. del debbio , a. di giacomo , b. lucini , nucl .
b594 ( 2001 ) 287 .
g. thooft , nucl .
b79 ( 1974 ) 276 . l. del debbio , a. di giacomo , g. paffuti , p. pieri , phys . lett .
b355 ( 1995 ) 255 .
a. di giacomo , , b. lucini , l. montesi , g. paffuti , phys .
d61 ( 2000 ) 034500 .
a. di giacomo , , b. lucini , l. montesi , g. paffuti , phys .
d61 ( 2000 ) 034505 . j. m. carmona , m. delia , a. di giacomo , , b. lucini , g. paffuti , phys . rev .
d64 ( 2001 ) 114507 .
e. marino , b. schroer , j. a. swieca , nucl .
b200 ( 1982 ) 473 . a. di giacomo , g.paffuti , 19th international symposium on lattice field theory , berlin 2001 , hep - lat 0110061 , to appear in the proceedings
. j. m. carmona , a. di giacomo , , b. lucini , phys .
b485 ( 2000 ) 126 . l. del debbio , a. di giacomo , g. paffuti , phys . lett .
b349 ( 1995 ) 513 . j. carmona , m. delia , l. del debbio , a. di giacomo , b. lucini , g. paffuti , contribution to lattice 01 berlin , hep - lat 010058 . l. del debbio , a. di giacomo , s. betti , in preparation | investigations on the structure of qcd vacuum from first principles can be done on the lattice .
the mechanism of confinement is an example : results from lattice on it are reviewed . |
a diagram @xmath0 of a knot @xmath1 is a nice representative of the knot type @xmath5 $ ] which is the isotopy class of @xmath1 .
it can be obtained from a regular planar projection ( or simply , regular projection ) @xmath6 of @xmath1 as follows .
let us take a sufficiently small neighborhood of each double point of @xmath6 so that the intersection of the neighborhood and @xmath6 is of the ` shape x ' on the plane . then modify the interior of each neighborhood so that we get a knot @xmath0 which is isotopic to @xmath1 and regularly projected to @xmath6 . in this sense
, a knot diagram can be ` almost planar ' , i.e. , it lies in the plane except for a sufficiently small neighborhood of each double point of the regular planar projection .
tabulation of knots is usually done by listing knots using their diagrams of minimal numbers of crossings . in general ,
such a minimal knot diagram is not unique within its isotopy class .
can this nonuniqueness be fixed by considering some other quantities associated with knot diagrams ? in this paper , we study the number of overpasses of a minimal knot diagram .
an immediate concern in this investigation is the relationship between the number of crossings and the number of overpasses of a knot diagram .
as it turns out , the number of crossings can be estimated from below and above by that of overpasses if the number of crossings is minimal among all diagrams of the same knot type ( theorem 2.9 ) . as an interesting consequence of this estimation of the minimal crossing number by the number of overpasses
, we may put two nice and relatively well understood classes of knots , that of alternating knots and that of @xmath7-torus knots , on the opposite extremes of a measurement of nonalternatingness of knot diagrams ( theorem 2.10 ) . under this measurement
, we prove that the ` most nonalternating ' knot diagrams are exactly the standard minimal @xmath7-torus knot diagrams ( theorem 3.1 ) .
notice that the minimal number of overpasses of a knot was the classical _ bridge number _ of a knot , first studied by schubert in @xcite , where the effect of various operations on knots ( satellite , cabling , connected sum ) on this number was investigated . throughout this paper ,
all knots are oriented and lie in the 3-dimensional sphere @xmath8 .
also , all knots are tame , i.e. , they are isotopic to polygonal or smooth knots .
hence , every knot has a diagram whose number of crossings is finite . for convenience ,
we distinguish knot diagrams and knots . from now on ,
a knot means an isotopy knot type @xmath9 $ ] for some knot diagram @xmath0 .
we will refer the reader to textbooks of knot theory ( such as @xcite or @xcite ) for standard terminologies .
also , we assume the _ well ordering property _ of the set @xmath10 of nonnegative integers to guarantee existence of the smallest element of any nonempty subset of @xmath10 .
we would like to thank professor marek chrobak of the department of compute science and engineering at uc riverside for asking stimulating questions which led our inequality about the numbers @xmath2 and @xmath3 to its current optimal form .
the number of crossings of a knot diagram @xmath0 is denoted by @xmath2 . for each knot @xmath1 , we denote @xmath11 _ is a diagram of _ @xmath12 by @xmath13
. note that we may assume a knot diagram @xmath0 lies in the plane by indicating ` overcrossings ' and ` undercrossings ' .
a ` crossing ' , in normal sense , of a knot diagram @xmath0 means a ` double point ' of the regular projection of @xmath0 .
hence , @xmath2 is the number of all double points of the regular projection of @xmath0 .
if @xmath14 is a crossing of a knot diagram @xmath0 , we denote the overcrossing and the undercrossing of @xmath0 projected to @xmath14 by @xmath15 and @xmath16 , respectively , where @xmath15 is above @xmath14 and @xmath16 is below @xmath14 .
alternatively , we may regard a crossing of @xmath0 as a pair of two points , overcrossing and undercrossing , in @xmath0 which are projected to the same double point . in this case , a crossing is considered as the preimage of a double point under the projection map .
let @xmath0 be a knot diagram with @xmath17 .
then there is a unique positive integer @xmath18 such that there is a finite sequence @xmath19 of @xmath20 points of @xmath0 , each of which is neither an overcrossing nor an undercrossing of @xmath0 , such that @xmath21,[s_2,f_2], ... ,[s_{k-1},f_{k-1}],[s_k , f_k]\ ] ] and @xmath22,[f_2,s_3], ...
,[f_{k-1},s_k],[f_k , s_1]\ ] ] are the _ overpasses _ and the _ underpasses _ of @xmath0 with respect to @xmath23 , respectively , where @xmath24 $ ] , for each @xmath25 , is the closed arc of @xmath0 from @xmath26 to @xmath27 which contains at least one overcrossing but has no undercrossing ; similarly , @xmath28 $ ] , for each @xmath29 , with the subscript counted modulo @xmath18 , is the closed arc of @xmath0 from @xmath27 to @xmath30 which contains at least one undercrossing but has no overcrossing .
we may assume @xmath31 .
if @xmath32 , then @xmath0 is of the ` shape 8 ' , and hence , @xmath33 .
existence : we will construct a finite sequence as stated above .
the following procedure is one way to get it .
let us start at a point @xmath34 on @xmath0 slightly before an undercrossing and go along with @xmath0 until arriving at the first overcrossing @xmath35 from @xmath34 .
let @xmath36 be the undercrossing just before @xmath35 .
take a point @xmath37 of @xmath0 between @xmath36 and @xmath35 .
then go along with @xmath0 from @xmath37 until arriving at the first undercrossing @xmath38 from @xmath37 , and let @xmath39 be the overcrossing just before @xmath38 .
take a point @xmath40 of @xmath0 between @xmath39 and @xmath38 .
we may now repeat this procedure to take the other points @xmath41 until there is no overcrossing between @xmath42 and @xmath34 . since the number @xmath2 of crossings of @xmath0 is finite
, @xmath18 must be finite . by its construction ,
the sequence of points @xmath23 of @xmath0 is a one as stated in the proposition .
uniqueness : suppose that @xmath43 is a positive integer and @xmath44 is a sequence of @xmath0 as stated in the proposition .
let @xmath45 be the sequence of points on @xmath0 as described above
. then @xmath46 must be contained in an arc of @xmath0 which contains @xmath26 for some @xmath25 but no overcrossing and no undercrossing of @xmath0 . in other words
, @xmath46 must be between @xmath47 and @xmath48 for some @xmath25 .
hence , @xmath49 must be between @xmath50 and @xmath51
. keep going on like this .
if @xmath52 , then @xmath53 must be between @xmath54 and @xmath55 , and if @xmath56 , then @xmath53 must be between @xmath57 and @xmath58 .
hence , we have @xmath59 .
this proves the uniqueness of @xmath18 .
such a sequence @xmath23 as in proposition 2.1 is said to be an _ over - underpass sequence _ of @xmath0 .
since any over - underpass sequence of @xmath0 consists of @xmath20 points , the number of overpasses ( or underpasses ) with respect to any over - underpass sequence of @xmath0 is @xmath18 .
hence , we define the number of overpasses ( or underpasses ) of the knot diagram @xmath0 as @xmath18 , and denote it by @xmath3 , sometimes called the length of over - underpass sequence . if a knot diagram @xmath0 has no crossing , we define @xmath3 as 0 . also , for each knot @xmath1 , we denote @xmath60 _
is a diagram of
_ @xmath12 by @xmath61
. this number @xmath61 is called the _ bridge number _ of @xmath1 . notice that ( 1 ) @xmath2 and @xmath3 are plane isotopy invariants of knot diagrams , i.e. , if @xmath62 and @xmath63 are plane isotopic knot diagrams , then @xmath64 and @xmath65 ; ( 2 ) @xmath13 and @xmath61 are isotopy invariants of knots , i.e. , if @xmath66 and @xmath67 are isotopic knots , then @xmath68 and @xmath69 . if @xmath70 , then @xmath0 is a diagram of a trivial knot . therefore , a knot @xmath1 is trivial if and only if @xmath1 has a diagram @xmath0 with @xmath70 .
obviously , for any positive integer @xmath18 , there is a diagram of a trivial knot whose number of overpasses is greater than @xmath18 . on the other hand ,
given a knot diagram @xmath0 with at least one crossing , we can add crossings to @xmath0 as many as we want without changing the knot type and the number of overpasses of @xmath0 .
for example , take a sufficiently small arc of @xmath0 from @xmath37 to a point between @xmath37 and the first overcrossing of @xmath0 from @xmath37 , and twist it alternatively so that the number of overpasses of @xmath0 is not changed . or
, we may modify the interior of a sufficiently small neighborhood of a crossing of @xmath0 to achieve the goal of increasing @xmath2 arbitrarily while keeping @xmath3 and the knot type of @xmath0 fixed .
hence , we have the following corollary . if @xmath0 is a diagram of a knot @xmath1 such that @xmath71 , then for every positive integer @xmath72 , there is a diagram @xmath73 of @xmath1 such that @xmath74 and @xmath75 .
remark that the number of crossings of a knot diagram with a minimal number of overpasses can be arbitrarily large .
@xmath76 for any knot diagram @xmath0 .
the equality holds if and only if @xmath0 is an alternating knot diagram .
furthermore , @xmath77 for any knot @xmath1 .
suppose that @xmath78 and @xmath23 is an over - underpass sequence of @xmath0 .
let @xmath79 be the number of overcrossings of @xmath0 on @xmath80 $ ] for each @xmath25 , @xmath81 the number of undercrossings of @xmath0 on @xmath28 $ ] for each @xmath82 , and @xmath83 the number of undercrossings of @xmath0 on @xmath84 $ ] .
then @xmath85 @xmath0 is an alternating knot diagram if and only if every overpass and underpass has exactly one overcrossing and undercrossing , respectively , i.e. , @xmath86 for each @xmath25 .
also , it follows immediately that @xmath77 for any knot @xmath1 .
the following lemma is the first key to prove the first main theorem of this paper ( theorem 2.9 ) . by this lemma , if an overpass of a knot diagram @xmath0 with respect to an over - underpass sequence of @xmath0 crosses an underpass more than once , then the number of crossings of @xmath0 is not minimal any more .
if @xmath0 is a minimal diagram of a knot @xmath1 with respect to crossings , i.e. , @xmath87 , @xmath3 is a nonnegative integer @xmath18 , and @xmath23 is an over - underpass sequence of @xmath0 , then every overpass of @xmath0 with respect to @xmath23 crosses each underpass at most once .
we may assume @xmath1 is nontrivial .
if @xmath1 is trivial , then @xmath88 , hence , @xmath89 . let @xmath0 be a diagram of a knot @xmath1 such that @xmath87 .
suppose that @xmath78 and @xmath23 is an over - underpass sequence of @xmath0 .
then , by lemma 2.4 , @xmath90 .
let @xmath91 $ ] and @xmath92 $ ] for each @xmath93 , and let @xmath94 $ ] and @xmath95 $ ] .
suppose that the number of crossings between @xmath96 and @xmath97 is at least 2 for some @xmath98 .
let @xmath99 and @xmath100 be two crossings of @xmath0 such that @xmath101 and @xmath102 are the first and the second overcrossings from @xmath26 among all overcrossings between @xmath96 and @xmath97 , respectively , and let @xmath103 and @xmath104 be two crossings of @xmath0 such that @xmath105 and @xmath106 are the first and the second undercrossings from @xmath107 in @xmath108 , respectively .
then either @xmath109 , @xmath110 or @xmath111 , @xmath112 .
there are two different diagrams for each case .
one of them for the case of @xmath109 , @xmath110 is shown in figure 1 .
it should be not hard for the reader to figure out the other diagrams .
let @xmath113 be the number of overcrossings on the arc @xmath114 of @xmath96 from @xmath101 to @xmath102 , and let @xmath115 be the number of undercrossings on the arc @xmath116 of @xmath97 from @xmath105 to @xmath106 . to prove this lemma , for each of the cases @xmath117 and @xmath118 , we will construct a diagram @xmath73 of @xmath1 by modifying an arc of @xmath96 or @xmath97 such that @xmath119 .
temporarily , we regard the knot diagram @xmath0 as its regular projection . hence , @xmath0 lies in the plane , and all overcrossings and all undercrossings of @xmath0 are the double points of @xmath0 as the regular projection .
consider the @xmath113 double points @xmath120 on the arc @xmath121 of @xmath96 from @xmath99 to @xmath100 such that @xmath122 with respect to the order from @xmath26 and the @xmath115 double points @xmath123 on the arc @xmath124 of @xmath97 from @xmath103 to @xmath104 such that @xmath125 with respect to the order from @xmath107 .
then @xmath126 , @xmath127 and @xmath128 , @xmath129 .
notice that we can take a sufficiently small positive real number @xmath130 , an @xmath130-neighborhood @xmath131 of @xmath121 , and an @xmath130-neighborhood @xmath132 of @xmath124 such that @xmath133 , @xmath134 , the set of all double points of @xmath0 contained in @xmath135 is @xmath136 , the set of all double points of @xmath0 contained in @xmath137 is @xmath138 , and , for every positive real number @xmath139 , @xmath140 and @xmath141 , where @xmath142 and @xmath143 are the boundaries of @xmath144 and @xmath145 , respectively .
hence , we may assume that @xmath146 and @xmath147 , where @xmath35 and @xmath36 are the first and the second points from @xmath26 in @xmath148 , respectively , @xmath149 and @xmath150 are the first and the second points from @xmath107 in @xmath151 , respectively , @xmath152 is an arc of the underpass passing through @xmath153 whose endpoints are on @xmath154 for each @xmath155 , and @xmath156 is an arc of the overpass passing through @xmath157 whose endpoints are on @xmath158 for each @xmath159 .
remark that @xmath160 is pairwise disjoint , @xmath161 and @xmath162 for each @xmath163 , and @xmath164 ; similarly , @xmath165 is pairwise disjoint , @xmath166 and @xmath167 for each @xmath168 , and @xmath169 .
let @xmath170 , @xmath171 , @xmath172 , @xmath173 be the first , the second , the third , the fourth points from @xmath107 in @xmath174 , respectively , and let @xmath175 be a point on @xmath176 between @xmath170 and @xmath103 . draw an arc @xmath177 in @xmath131 starting at @xmath175 so that @xmath177 intersects @xmath152 transversely only once for each @xmath178 but @xmath177 does not intersect @xmath96 .
then we take @xmath179 as the intersecting point of @xmath177 and @xmath180 and denote the arc of @xmath177 from @xmath175 to @xmath179 by @xmath181 .
similarly , let @xmath182 , @xmath183 , @xmath184 , @xmath185 be the first , the second , the third , the fourth points from @xmath26 in @xmath186 , respectively , and let @xmath187 be a point on @xmath188 between @xmath182 and @xmath99 . draw an arc @xmath189 in @xmath132 starting at @xmath187 so that @xmath189 intersects @xmath156 transversely only once for each @xmath190 but @xmath189 does not intersect @xmath97 .
then we take @xmath191 as the intersecting point of @xmath189 and @xmath192 and denote the arc of @xmath189 from @xmath187 to @xmath191 by @xmath193 . from now on
, @xmath0 is the diagram of @xmath1 again , i.e. , @xmath0 is not a regular projection of a knot but a knot diagram .
let @xmath194 be a regularly projecting arc in @xmath195 whose regular projection is @xmath181 such that @xmath194 has no overcrossing , and let @xmath196 be a regularly projecting arc in @xmath195 whose regular projection is @xmath193 such that @xmath196 has no undercrossing .
then figure 1 .
proof of lemma 2.5 : case 1.(1 ) and case 2.(2 ) when @xmath109 , @xmath110 .
@xmath117 : let @xmath197 .
then @xmath73 is a diagram of @xmath1 .
\(1 ) if @xmath176 and @xmath180 make the same sign with @xmath96 , then @xmath198 , hence , @xmath199 ; \(2 ) if @xmath176 and @xmath180 make opposite signs with @xmath96 , then @xmath200 , hence , @xmath201 .
@xmath118 : let @xmath202 .
then @xmath73 is a diagram of @xmath1 .
\(1 ) if @xmath188 and @xmath192 make the same sign with @xmath97 , then @xmath203 , hence , @xmath204 ; \(2 ) if @xmath188 and @xmath192 make opposite signs with @xmath97 , then @xmath205 , hence , @xmath206 .
therefore , we have a diagram @xmath73 of @xmath1 such that @xmath119 .
this is a contradiction to @xmath87 .
now , we prove the second key lemma for theorem 2.9 by a similar idea as the one used in the proof of lemma 2.5 . by this lemma , if an overpass of a knot diagram @xmath0 with respect to an over - underpass sequence of @xmath0 crosses an adjacent underpass or an underpass crosses an adjacent overpass , then the number of crossings of @xmath0 is not minimal .
if @xmath0 is a minimal diagram of a knot @xmath1 with respect to crossings , @xmath207 , and @xmath23 is an over - underpass sequence of @xmath0 , then no overpass of @xmath0 with respect to @xmath23 crosses its adjacent underpasses and no underpass crosses its adjacent overpasses , where the adjacent underpasses of the @xmath208-th overpass @xmath96 are the @xmath209-th and the @xmath208-th underpasses @xmath210 and @xmath211 for each @xmath212 ; the adjacent underpasses of @xmath213 are @xmath214 and @xmath215 ; the adjacent overpasses of @xmath211 are @xmath96 and @xmath216 for each @xmath93 ; the adjacent overpasses of @xmath214 are @xmath217 and @xmath213 .
we may assume that each overpass of @xmath0 with respect to @xmath23 crosses each underpass at most once by lemma 2.5 .
suppose that there is @xmath25 such that the @xmath208-th overpass @xmath91 $ ] of @xmath0 with respect to @xmath23 crosses the @xmath208-th underpass @xmath92 $ ] , with the subscript counted modulo @xmath18 .
let @xmath14 be the crossing of @xmath0 between @xmath96 and @xmath211 .
let @xmath218 be the crossing of @xmath0 such that @xmath219 is the overcrossing of @xmath96 just before @xmath27 , and let @xmath220 be the crossing of @xmath0 such that @xmath221 is the undercrossing of @xmath211 just after @xmath27 .
note that @xmath14 , @xmath218 , @xmath220 need not be distinct .
however , if some of them are identical , then , by reidemeister moves , we can reduce the crossing easily .
let @xmath113 be the number of overcrossings on the arc @xmath222 of @xmath96 from @xmath15 to @xmath219 , and let @xmath115 be the number of undercrossings on the arc @xmath223 of @xmath211 from @xmath221 to @xmath16 .
temporarily , we regard the knot diagram @xmath0 as its regular projection as we did in the the proof of lemma 2.5 .
let us take an @xmath130-neighborhood @xmath131 of @xmath224 and an @xmath130-neighborhood @xmath225 of @xmath226 for a sufficiently small positive real number @xmath130 as described in the proof of lemma 2.5 .
let @xmath35 and @xmath36 be the first and the second points from @xmath26 in @xmath227 , respectively , and let @xmath38 and @xmath39 be the first and the second points from @xmath26 in @xmath228 , respectively .
similarly , let @xmath149 and @xmath150 be the first and the second points from @xmath26 in @xmath229 , respectively , and let @xmath230 and @xmath231 be the first and the second points from @xmath26 in @xmath232 , respectively .
let @xmath233 be the arc of @xmath234 from @xmath36 to @xmath39 such that @xmath235 , and let @xmath236 be the arc of @xmath237 from @xmath149 to @xmath230 such that @xmath238 . from now on
, @xmath0 is the diagram of @xmath1 again .
let @xmath239 be a regularly projecting arc in @xmath195 whose regular projection is @xmath233 such that @xmath239 has no overcrossing , and let @xmath240 be a regularly projecting arc in @xmath195 whose regular projection is @xmath236 such that @xmath240 has no undercrossing .
then case 1 .
@xmath117 : let @xmath241 .
then @xmath73 is a diagram of @xmath1 and @xmath242 , hence , @xmath199 .
@xmath118 : let @xmath243 .
then @xmath73 is a diagram of @xmath1 and @xmath244 , hence , @xmath204 .
therefore , we have a diagram @xmath73 of @xmath1 such that @xmath119 .
this is a contradiction to @xmath87 .
= 3.5 in figure 2 .
proof of lemma 2.6 . as the other case , if there is @xmath25 such that the @xmath208-th underpass @xmath92 $ ] of @xmath0 with respect to @xmath23 crosses the @xmath245-th overpass @xmath246 $ ] , with the subscript counted modulo @xmath18 , then we can construct a diagram @xmath73 of @xmath1 such that @xmath119 by the same argument above .
this proves the lemma .
the following two lemmas are immediate consequences of lemma 2.6 and the other keys for theorem 2.9 .
if @xmath0 is a minimal diagram of a knot @xmath1 with respect to crossings , @xmath3 is a positive integer @xmath18 , and @xmath23 is an over - underpass sequence of @xmath0 , then there is no overpass of @xmath0 with respect to @xmath23 which crosses all underpasses .
suppose that an overpass of @xmath0 with respect to @xmath23 crosses all underpasses .
then the overpass crosses its adjacent underpasses .
this is a contradiction to @xmath87 by lemma 2.6 .
if @xmath0 is a minimal diagram of a knot @xmath1 with respect to crossings , @xmath207 , and @xmath23 is an over - underpass sequence of @xmath0 , then every overpass of @xmath0 with respect to @xmath23 crosses at most @xmath247 underpasses .
suppose that an overpass @xmath248 of @xmath0 with respect to @xmath23 crosses @xmath249 underpasses .
then , by lemma 2.5 , @xmath248 crosses each of the @xmath249 underpasses exactly once . since every overpass has two adjacent underpasses and @xmath248 crosses @xmath249 underpasses , @xmath248 must cross at least one of its adjacent underpasses
this is a contradiction to @xmath87 by lemma 2.6 . if @xmath0 is a minimal diagram of a knot @xmath1 with respect to crossings , then @xmath250 .
let @xmath0 be a diagram of a knot @xmath1 such that @xmath87 . if @xmath1 is a trivial knot , then @xmath0 is a simple closed curve on the plane , hence , @xmath251 .
remark that @xmath3 can not be @xmath252 . if @xmath253 , then , by corollary 2.2 , @xmath0 represents a trivial knot , i.e. , @xmath1 is a trivial knot , hence , @xmath88 .
this contradicts to lemma 2.4 .
next , we claim that @xmath3 can not be 2 if @xmath0 is a minimal diagram of a nontrivial knot @xmath1 with respect to crossings .
this follows from the same argument in the proof of lemma 2.8 .
suppose that @xmath1 is a nontrivial knot , @xmath87 , @xmath254 , and @xmath255 is an over - underpass sequence of @xmath0 .
then each overpass of @xmath0 with respect to @xmath255 must cross one of the underpasses .
however , all underpasses are adjacent to each overpass .
suppose that @xmath256 and @xmath23 is an over - underpass sequence of @xmath0 .
we may assume that the overpasses @xmath257,[s_2,f_2], ... ,[s_{k-1 } , f_{k-1}],[s_k , f_k]$ ] are disjoint closed arcs on the projection plane of @xmath0 and the underpasses @xmath258,[f_2,s_3], ... ,[f_{k-1},s_k],[f_k , s_1]$ ] are also disjoint closed arcs on it .
remark that the number of crossings of @xmath0 is the number of intersections of projections of overpasses and underpasses of @xmath0 .
since we have @xmath18 overpasses and @xmath18 underpasses , by lemma 2.5 and lemma 2.8 , there are at most @xmath259 crossings among overpasses and underpasses . notice that @xmath260 does not hold in general .
an example can be given with @xmath1 being the @xmath261-torus knot .
we have @xmath262 in this case .
the following theorem is an immediate consequence of theorem 2.9 if @xmath1 is nontrivial .
it means that the number of overpasses of a minimal knot diagram with respect to crossings can also be estimated by the number of crossings .
once more , in this case , the number of overpasses needs not be minimal and is at least 3 as shown in the proof of theorem 2.9 .
if @xmath1 is a nontrivial knot and @xmath0 is a minimal diagram of @xmath1 with respect to crossings , then @xmath263 .
theorem 2.10 provides us the most optimal lower bound for the number of overpasses of @xmath0 when @xmath0 is a minimal knot diagram with respect to crossings . here , an interesting problem occurs . in lemma 2.4
, we showed that a knot diagram @xmath0 is alternating if and only if @xmath264 .
also , by kauffman @xcite , an alternating knot has a minimal knot diagram with respect to crossings which is alternating .
hence , the second inequality of theorem 2.10 becomes an equality if and only if the knot is an alternating knot . on the other hand , as shown by murasugi @xcite , the standard diagram @xmath0 of a @xmath7-torus knot is a minimal knot diagram with respect to crossings and @xmath265 .
so the first inequality of theorem 2.10 becomes an equality for the @xmath7-torus knot with @xmath266 .
is the converse true under the condition that @xmath0 is a minimal knot diagram with respect to crossings ? we prove this as the last theorem of this paper .
if @xmath0 is a minimal diagram of a nontrivial knot @xmath1 with respect to crossings and @xmath267 , then @xmath0 is the standard diagram of either the @xmath268-torus knot or the @xmath269-torus knot .
hence , @xmath1 is the @xmath270-torus knot .
let @xmath0 be a minimal diagram of a nontrivial knot @xmath1 with respect to crossings such that @xmath267 .
suppose that @xmath266 .
then @xmath271 ( see the proof of theorem 2.9 ) .
in order to prove this theorem , we will consider all possible knot diagrams satisfying the hypothesis and show that they can only be the standard diagrams of either the @xmath272-torus knot or the @xmath273-torus knot depending on signs of crossings .
for convenience , we regard the knot diagram @xmath0 as its regular projection here . also , to visualize overpasses and underpasses clearly , we imagine blue and red colors for overpasses and underpasses , respectively . in this sense , crossings are only the intersections of blue arcs and red arcs on the plane , i.e. , crossings are purple !
now , let us start drawing all possible knot diagrams satisfying the hypothesis .
notice that such knot diagrams must satisfy the following 3 rules : rule 1 : every overpass and every underpass intersect each underpass and each overpass at most once , respectively ( by lemma 2.5 ) ; rule 2 : no overpass and no underpass intersect its adjacent underpasses and its adjacent overpasses , respectively ( by lemma 2.6 ) ; rule 3 : every overpass and every underpass intersect @xmath247 underpasses and @xmath247 overpasses , respectively ( by lemma 2.8 ) .
if some overpass has less than @xmath247 overcrossings , then at least one overpass has more than @xmath247 overcrossings ; similarly , if some underpass has less than @xmath247 undercrossings , then at least one underpass has more than @xmath247 undercrossings .
this is a contradiction to lemma 2.8 .
from now on , we describe drawing the knot diagrams .
let us take a point @xmath37 on the plane and draw the first overpass @xmath274 $ ] .
draw the first underpass @xmath275 $ ] and the second overpass @xmath276 $ ] . until now
, we can not have any crossing by rule 2 .
when we draw the second underpass @xmath277 from @xmath278 , we must make @xmath277 intersect @xmath213 first by rule 2,3 .
notice that we have only 2 cases that @xmath277 intersects @xmath213 first as follows .
the sign @xmath279 of crossing between @xmath277 and @xmath213 is @xmath280 , where @xmath281 if @xmath213 intersects @xmath277 from left to right and @xmath282 if @xmath213 intersects @xmath277 from right to left when we look @xmath277 as a line segment passing through the crossing upward .
notice that , by rule 1 , we can define the sign of each crossing by this way .
after @xmath277 intersects @xmath213 first with @xmath283 , we must change color from red to blue by rule 1,2 .
take a point @xmath284 so that @xmath285 $ ] has only one undercrossing .
then the third overpass @xmath286 must intersect @xmath215 first with @xmath287 by rule 2,3 .
after @xmath286 intersects @xmath215 first , we must change color from blue to red by rule 1,2 .
take a point @xmath288 so that @xmath289 $ ] has only one overcrossing .
then the third underpass @xmath290 can intersect only either @xmath213 first with @xmath291 or @xmath292 first with @xmath293 .
we claim that @xmath290 must intersect @xmath292 first with @xmath293 .
suppose that @xmath290 intersects @xmath213 first with @xmath287 .
then , after @xmath290 intersects @xmath213 first , we must change color from red to blue by rule 1,2 .
take a point @xmath294 so that @xmath295 $ ] has only one undercrossing . in this case
, we can not draw any knot diagram such that @xmath290 intersects @xmath292 .
this is a contradiction to rule 2,3 .
hence , @xmath290 intersects @xmath292 first with @xmath293 .
after @xmath290 intersects @xmath292 first , take a point @xmath296 which is neither @xmath37 nor a crossing so that the arc @xmath297 $ ] of @xmath290 has only one undercrossing .
notice that we can connect @xmath296 and @xmath37 by an arc without crossing .
= 4.8 in figure 3 .
the first three overpasses and underpasses , @xmath283 .
if we have only 3 overpasses and 3 underpasses , we must connect @xmath296 and @xmath37 by an arc without crossing .
this is the only way to draw the knot diagrams satisfying rule 1,2,3 and gives us the standard diagram of the @xmath298-torus knot .
see figure 3 , where solid arcs are supposed to be of the blue color ( overpasses ) and dashed arcs of the red color ( underpasses ) .
suppose that we have more than 3 overpasses , i.e. , @xmath299 .
then the third underpass @xmath290 must intersect @xmath213 second by rule 1,2,3 and we have only 2 cases that @xmath290 intersects @xmath213 second . * claim : * @xmath290 must intersect @xmath213 second with @xmath291 .
see figure 4(i ) .
_ proof of the claim : _ suppose that @xmath290 intersects @xmath213 second with @xmath300 . see figure 4(ii ) .
then , after @xmath290 intersects @xmath213 second , we must change color from red to blue by rule 1 .
take a point @xmath294 so that @xmath295 $ ] has only two undercrossings .
then @xmath301 must intersect @xmath277 first with @xmath302 and @xmath215 second with @xmath303 by rule 1,2,3 . see figure 4(iii ) .
after @xmath301 intersects @xmath215 second , we must change color from blue to red by rule 1 .
take a point @xmath304 so that @xmath305 $ ] has only two overcrossings .
then @xmath306 can intersect only either @xmath213 first with @xmath307 or @xmath286 first with @xmath308 .
however , @xmath306 can not intersect @xmath213 first .
if @xmath306 intersects @xmath213 first with @xmath307 , we must change color from red to blue by rule 1,2 . by a similar argument
as the one used before , in this case , we can not draw any knot diagram such that @xmath306 intersects @xmath292 .
this is a contradiction to rule 2,3 .
hence , @xmath306 must intersect @xmath286 first with @xmath308 .
after @xmath306 intersects @xmath286 first , @xmath306 can intersect only either @xmath213 second with @xmath307 or @xmath292 second with @xmath309 .
however , @xmath306 can not intersect @xmath213 second .
if @xmath306 intersects @xmath213 second with @xmath307 , after @xmath306 intersects @xmath213 second , we must change color from red to blue by rule 1,2 . in this case
, we can not draw any knot diagram such that @xmath306 intersects @xmath292 .
this is a contradiction to rule 2,3 .
hence , the only way to draw @xmath306 is that @xmath306 intersects @xmath286 first with @xmath308 and @xmath292 second with @xmath309 .
see figure 4(iv ) .
= 4.5 in figure 4 .
the first three overpasses and underpasses can only be drawn as in ( i ) .
after @xmath306 intersects @xmath292 second , we must change color from red to blue by rule 1 .
take a point @xmath310 so that @xmath311 $ ] has only two undercrossings .
then @xmath312 can intersect only either @xmath215 first with @xmath313 or @xmath290 first with @xmath314 .
however , @xmath312 can not intersect @xmath215 first .
if @xmath312 intersects @xmath215 first with @xmath313 , then we must change color from blue to red by rule 1,2 . in this case
, we can not draw any knot diagram such that @xmath312 intersects @xmath277 .
this is a contradiction to rule 2,3 .
hence , @xmath312 must intersect @xmath290 first with @xmath314 . again , see figure 4(iv ) .
after @xmath312 intersects @xmath290 first , @xmath312 can intersect only either @xmath215 second with @xmath313 or @xmath277 second with @xmath315 .
however , @xmath312 can not intersect @xmath215 second .
if @xmath312 intersects @xmath215 second with @xmath313 , after @xmath312 intersects @xmath215 second , we must change color from blue to red by rule 1,2 . in this case
, we can not draw any knot diagram such that @xmath312 intersects @xmath277 .
this is a contradiction to rule 2,3 .
hence , @xmath312 must intersect @xmath277 second with @xmath315 .
once more , see figure 4(iv ) .
after @xmath312 intersects @xmath277 second , we must change color from blue to red by rule 1 . in this case
, we can not draw any knot diagram such that @xmath312 intersects @xmath215 .
this is a contradiction to rule 2,3 .
hence , when @xmath300 , we can not draw any knot diagram satisfying rule 1,2,3 . therefore , @xmath316 must be @xmath280 .
this finishes the proof of the claim .
now , we use an induction on the order of the over - underpass sequence .
the argument used here is just a generalization of the previous one .
hence , our argument will be slightly sketchy here .
suppose , without loss of generality , that @xmath317 and @xmath318 is an over - underpass sequence of the standard diagram of the @xmath319-torus knot .
= 5.5 in figure 5 .
it is impossible for @xmath320 to intersect @xmath213 with @xmath321 .
let @xmath91 $ ] and @xmath92 $ ] for each @xmath322 , and let @xmath323 $ ] and @xmath324 $ ] .
take a point @xmath325 of @xmath326 between the @xmath327-th undercrossing of @xmath326 and @xmath37 .
then the arc @xmath328 $ ] of @xmath326 has also @xmath329 undercrossings .
notice that @xmath330 are overpasses and underpasses of a knot diagram @xmath0 satisfying rule 1,2,3 and @xmath328 $ ] is an arc of the @xmath72-th underpass @xmath320 of @xmath0 .
suppose that @xmath0 has more than @xmath72 overpasses .
then there is an @xmath331-th undercrossing on @xmath320 by @xmath213 according to rule 1,2,3 .
we claim that at this undercrossing , @xmath320 must intersect @xmath213 with @xmath321 .
to prove this claim , suppose that @xmath320 intersects @xmath213 with @xmath332 .
see figure 5 . then
, after @xmath320 intersects @xmath213 with @xmath332 , we must change color from red to blue by rule 1 .
take a point @xmath333 so that @xmath334 $ ] has @xmath335 undercrossings .
then the only way to draw @xmath336 is that @xmath336 intersects @xmath337 first , @xmath338 second , ... , @xmath277 @xmath327-th , and @xmath215
@xmath331-th so that @xmath339 , and @xmath340 by rule 1,2,3 .
after @xmath336 intersects @xmath215 with @xmath340 , we must change color from blue to red by rule 1 .
take a point @xmath341 so that @xmath336 has @xmath335 overcrossings .
then the only way to draw @xmath342 is that @xmath342 intersects @xmath343 first , @xmath344 second , ... , @xmath286 @xmath327-th , and @xmath292 @xmath331-th so that @xmath345 , and @xmath346 by rule 1,2,3 . after @xmath342 intersects @xmath292 with @xmath346 , we must change color from red to blue by rule 1 .
take a point @xmath347 so that @xmath342 has @xmath335 undercrossings .
then the only way to draw @xmath348 is that @xmath348 intersects @xmath320 first , @xmath337 second , ... , @xmath290 @xmath327-th , and @xmath277 @xmath331-th so that @xmath349 , and @xmath350 by rule 1,2,3 . after @xmath348 intersects @xmath277 with @xmath350 , we must change color from blue to red by rule 1 .
however , in this case , we can not draw any knot diagram such that @xmath348 intersects @xmath215 .
this is a contradiction to rule 2,3 .
therefore , @xmath351 can not be @xmath352 . as a next step ,
we claim that the only way to draw @xmath336 and @xmath342 gives us the standard diagram of the @xmath353-torus knot ( see figure 6 ) .
this will finish the induction .
after @xmath320 intersects @xmath213 @xmath331-th with @xmath321 , we must change color from red to blue by rule 1,2 .
take a point @xmath333 so that @xmath334 $ ] has only @xmath335 undercrossings .
then the only way to draw @xmath336 is that @xmath336 intersects @xmath337 first , @xmath338 second , ... , @xmath215 @xmath331-th so that @xmath354 by rule 1,2,3 .
suppose that @xmath336 intersects @xmath211 first for some @xmath355
. then @xmath356 must be @xmath280 and @xmath336 must intersect @xmath211 first , @xmath210 second , ... , @xmath215 @xmath208-th so that @xmath357 by rule 1,2,3
. however , after @xmath336 intersects @xmath215 @xmath208-th with @xmath358 , @xmath336 must stop before intersecting @xmath320 and we must change color from blue to red by rule 1,2 . in this case
, we can not draw any knot diagram such that @xmath336 intersects @xmath337 .
this is a contradiction to rule 2,3 .
hence , @xmath336 must intersect @xmath337 first , @xmath338 second , ... , @xmath215 @xmath331-th so that @xmath354 . take a point @xmath341 so that @xmath359 $ ] has only @xmath335 overcrossings . then , by a similar argument as before , we can show that the only way to draw @xmath342 is that @xmath342 intersects @xmath343 first , @xmath344 second , ... , @xmath292 @xmath331-th so that @xmath360 by rule 1,2,3 .
the only way to draw @xmath336 and @xmath342 .
suppose that @xmath342 intersects @xmath96 first for some @xmath361
. then @xmath362 must be @xmath280 and @xmath342 must intersect @xmath96 first , @xmath363 second , ...
, @xmath213 @xmath208-th so that @xmath364 by rule 1,2,3
. however , after @xmath342 intersects @xmath213 @xmath208-th with @xmath365 , @xmath342 must stop before intersecting @xmath336 and we must change color from red to blue by rule 1,2 . in this case , we can not draw any knot diagram such that @xmath342 intersects @xmath343 .
this is a contradiction to rule 2,3 .
hence , @xmath342 must intersect @xmath343 first , @xmath344 second , ... , @xmath292 @xmath331-th so that @xmath360 .
take a point @xmath366 so that the arc @xmath367 $ ] of @xmath342 has only @xmath335 undercrossings .
notice that , in this case , we can connect @xmath366 and @xmath37 by an arc without crossing so that we complete drawing the knot diagram .
this is the only way to draw a knot diagram with @xmath368 overpasses satisfying the rules 1,2,3 . to complete the argument ,
when @xmath369 , we can get the standard diagram @xmath0 of the @xmath273-torus knot and this is the only way to draw the knot diagrams satisfying rule 1,2,3 , and hence , the theorem is proved in case 1 .
the sign @xmath279 of crossing between @xmath277 and @xmath213 is @xmath352 . by the same argument
as used in case 1 , we can get only standard diagrams of the @xmath268-torus knot .
this proves the theorem .
one of goeritz s unknot diagrams .
one may view lemmas 2.5 and 2.6 from another perspective .
suppose @xmath0 is a knot diagram , not necessarily being minimal with respect to crossings .
if either an overpass interesects an underpass more than once or an overpass interesects its adjacent underpass , then lemmas 2.5 and 2.6 give a specific way to change the knot diagram @xmath0 by isotopy with the number of crossings reduced .
for such a knot diagram @xmath0 , we have @xmath370 when @xmath264 , @xmath0 is an alternating knot diagram . by theorem 3.1 ,
when @xmath267 , @xmath0 is the standard diagram of a @xmath371 torus knot .
thus , it is natural to wonder whether conditions ( 1 ) and ( 2 ) above are sufficient for a knot diagram of a prime knot being minimal with respect to crossings .
goeritz s unknot diagrams show that the answer to this question is negative .
see figure 7 .
are there any other necessary conditions on the overpasses and underpasses of a knot diagram for it being minimal with respect to crossings ? this will be the topic of our further investigation . | given a diagram @xmath0 of a knot @xmath1 , we consider the number @xmath2 of crossings and the number @xmath3 of overpasses of @xmath0 .
we show that , if @xmath0 is a diagram of a nontrivial knot @xmath1 whose number @xmath2 of crossings is minimal , then @xmath4 .
these inequalities are shape in the sense that the upper bound of @xmath3 is achieved by alternating knots and the lower bound of @xmath3 is achieved by torus knots .
the second inequality becomes an equality only when the knot is an alternating knot .
we prove that the first inequality becomes an equality only when the knot is a torus knot . |
a variety of experiments@xmath14 points out that the small coherence length , reduced carrier density , high transition temperature @xmath10 and marked anisotropy of cuprate superconductors cause strong enhancement of superconducting fluctuations ( sf ) .
in contrast to conventional superconductors , in cuprates the transition region is considerably smeared by sf which can be detected in a wide temperature range , up to 10 - 15 k. the formation of the fluctuation cooper pairs above @xmath10 results in the appearance of a langevin - type diamagnetic contribution to the magnetization @xmath15 , existing side by side to the paramagnetic contribution from fermionic carriers .
since the size of fluctuating pairs @xmath16 grows when @xmath17 approaches the transition temperature @xmath10 , @xmath18 should diverge near the transition for any small fixed magnetic field , being equal zero for @xmath19 on the other hand it is evident that very strong magnetic fields , comparable to @xmath20 must suppress sf .
therefore the isothermal magnetization curve @xmath21 has to exhibit an upturn .
this upturn can be quantitatively described in the framework of the exactly solvable , for any magnetic field , zero - dimensional model@xmath22 ( superconducting granula with the size @xmath23 @xmath16 ) or by means of cumbersome microscopic treatment accounting for the short - wavelength fluctuation contribution in the @xmath24 case.@xmath25 the experiments on conventional bcs superconductors show that the magnetization is quenched for fields as low as @xmath26 ( see ref.@xmath27 .
the value of the upturn field @xmath28 in the magnetization curves can be considered inversely proportional to the coherence length.@xmath29 this explains why in optimally doped high - temperature superconductors , the ginzburg - landau ( gl ) picture works pretty well . here
the coherence length is so short that the quenching of fluctuating magnetization on increasing the magnetic field has not yet been observed .
the fluctuating magnetization of layered superconductors in the vicinity of the transition temperature and for h@xmath30 when the contribution of short - wavelength fluctuations@xmath25 is negligible , can be theoretically described@xmath31 in the framework of the gl scheme with the lawrence - doniach hamiltonian.@xmath32 the fluctuating diamagnetism ( fd ) turns out to be a complicated nonlinear function of temperature and magnetic field and can not be factorized on these variables .
an important role in fd is played by the degree of anisotropy of the electronic spectrum .
all these aspect of fd have been found to occur in optimally doped ybco.@xmath33 also scaling arguments@xmath34 were found@xmath35 rather well obeyed in this compound . in underdoped ybco ,
instead , marked deviations from the behaviour expected in the framework of gl approaches have been detected .
a first qualitative claim in this regard goes back to kanoda et al.,@xmath36 who noticed that in oxygen deficient ybco the fd in small fields was enhanced . later on ,
novel features of fd in underdoped compounds have been reported @xmath37 .
in particular , in underdoped ybco at @xmath38 marked enhancement of the susceptibility for fixed @xmath39 in a field of 0.02 tesla was detected above @xmath40 , with magnetization curves strongly different from the ones in optimally doped ybco.@xmath41 magnetization curves have been subsequently reported @xmath42 in underdoped @xmath43 ( lasco ) . for the moment we only mention that the magnetization curves reported by carballeira et al.@xmath42 in underdoped lasco , although indicating field - affected fluctuation - induced diamagnetism , do not exhibit the upturn with the magnetic field as the ones in underdoped ybco that we will discuss later on
. finally , recent magnetization data@xmath44 as a function of temperature in @xmath45 single crystal ( with transition temperature in zero field @xmath46 ) indicate sf obeying to @xmath47 scaling conditions for @xmath48 tesla and turning to @xmath24 scaling for smaller fields . qualitative justifications of the anomalous diamagnetism in underdoped ybco have been tried@xmath49 , essentially based on the idea of charge inhomogeneities leading to non - percolating superconducting drops or on the extension of the theory by ovchinnikov et al.@xmath50 , where the anomalous diamagnetism is related to regions having local @xmath10 s higher than the resistive transition temperature .
the first theoretical study specifically aimed at the description of fd in underdoped ybco was undertaken by sewer and beck @xmath51 . in the framework of the lawrence - doniach model ,
these authors justify the temperature and field dependences of the magnetic susceptibility by taking into account the phase fluctuations of the order parameter , thus arriving to a layered xy - model for a liquid of vortices . in this paper
we address the problem of the fluctuating diamagnetism in the y@xmath0ca@xmath1ba@xmath52cu@xmath3o@xmath4 family and of its dependence from the number of holes , by reporting squid magnetization measurements in a series of samples .
preliminary results on overdoped compounds have been presented to a meeting and published elsewhere.@xmath53 since some differences in the magnetic behaviour of chain - ordered and chain - disordered ybco have been noticed@xmath54 , we also attained the underdoped regime by means of @xmath55 for @xmath56 substitution in ideally chain - empty @xmath57 , while for the overdoped regime the same heterovalent substitution was performed in chain - full @xmath58 .
the paper is organized as follows . in sect.ii experimental details and the majority of the experimental results
are reported .
the analysis of the data ( sect.iii ) is first tentatively carried out on the basis of an anisotropic free energy gl functional , within the gaussian approximation .
the inapplicability of such an approach for non - optimally doped compounds is stressed .
then the theory for phase fluctuations of the order parameter in layered liquid of vortices is revised , to properly take into account terms neglected in the previous formulation.@xmath51 in particular non - reversibility and relaxation effects of the magnetization are argued to support the picture of non - percolating , locally superconducting droplets above the resistive transition temperature , that we interpret as phase fluctuations of a non - zero order parameter below the local irreversibility temperature
. thus a comprehensive description of fd in the y@xmath0ca@xmath1ba@xmath52cu@xmath3o@xmath4 family is obtained , as it is summarized in sect .
the samples of chemical composition @xmath59 were prepared by solid state reactions of oxides and carbonates in flowing oxygen at 1000 k for about 100 hours .
x - ray diffractometry was used to check the presence of a single phase .
the oxygen stoichiometry was first estimated by thermogravimetry and energy dispersive spectrometry .
the samples were then oxygenated close to @xmath60 by annealing in oxygen atmosphere ( 25 atm ) at@xmath61450 k for about 100 hours or deoxygenated as much as possible close to @xmath62 , for about 100 hours in vacuum .
the final oxygen content turned out @xmath63 for overdoped ybco and @xmath64 for the underdoped samples , estimated with loss of mass measures . before the measurements ,
the samples stay at room temperature for about 1 week .
the resistive transition in samples of the same batch appeared very sharp , with moderate evidence of paraconductivity in a temperature range of 5 - 10 k above the transition .
after mixing the samples with epoxy resin , they were oriented by hardening in a strong magnetic field ( 9 tesla ) .
the orientation was tested by comparing the diamagnetic susceptibility for h//c with the one for h in the ab plane , where practically no enhancement of m@xmath65 was noted to occur . for samples used in previous works @xmath66 ,
the orientation was also tested by means of the @xmath67cu nmr line ( see ref .
magnetization measurements have been carried out in the oriented powders by means of a quantum design mpms - xl7 squid magnetometer .
measurements were performed also in optimally doped ybco in order to prove that the results in oriented powders do not significantly differ from the ones in single crystals .
the data already obtained by other authors@xmath68 were confirmed . in the following section
we will recall a few results of the studies in optimally doped compounds , when it is required for the comparison of our data in strongly underdoped or overdoped ybco : ca .
the transition temperatures @xmath69 have been estimated from the magnetization curves vs. @xmath17 at small fields ( 20 oersted ) , by extrapolating at m=0 the linear behavior of @xmath70 occurring below t@xmath71 , as shown in the insets in fig.1 .
the values of @xmath40 are collected in table i , where the numbers of holes @xmath72 , as evaluated from the expression @xmath73 , giving the parabolic behaviour@xmath74 of the phase diagram @xmath75 , are also reported .
it is noted that because of the enhanced fluctuating diamagnetism some uncertainty in the estimate of t@xmath71(0 ) is present , particularly in strongly underdoped samples .
this uncertainty does not affect the discussion given later on about the anomalous diamagnetism , which is detected in a temperature region well above t@xmath71(0 ) .
magnetization measurements at constant field h have been performed as a function of temperature , with @xmath76 @xmath77 @xmath78 . in general two contributions to the magnetization m were observed : a pauli - like , positive term @xmath79 , almost @xmath17-independent or only slightly increasing with decreasing temperature in the range @xmath80 from 200k down to about 100 k and a negative diamagnetic contribution @xmath81 arising on approaching @xmath10 .
this latter contribution was extracted by subtracting from m the value obtained by extrapolating for @xmath82 the curve @xmath79 vs. @xmath17 in @xmath80 , where @xmath83 is practically zero .
thus the possible slight temperature dependence of @xmath79 around @xmath10 was neglected in comparison to the much stronger diamagnetic term .
typical magnetization curves @xmath84for overdoped and underdoped samples are reported in fig.2 .
the enhancement of fd , in both regimes , is evidenced . in fig.3 some isothermal magnetization curves @xmath85 vs h , obtained by cooling in zero magnetic field ( zfc ) down to a certain temperature above @xmath40 ,
are shown . in fig 3c also
a few data for field cooled ( fc ) magnetization , to be discussed later on , are reported .
the generalization of the gl functional for layered superconductors [ lawrence - doniach ( ld ) functional @xmath86 in a perpendicular magnetic field can be written @xmath87 & = & { \sum_{l}}\int d^{2}r\left ( \alpha \left| \psi _ { l}\right| ^{2}+\frac{\beta } { 2}\left| \psi _ { l}\right| ^{4}+% \frac{h^2}{16\pi^2 m}\left| \left ( { \bf \nabla } _ { \shortparallel } -\frac{2ie% } { c\hbar } { \bf a}_{\shortparallel } \right ) \psi _ { l}\right| ^{2}\right .
\nonumber \\ & & \left .
+ { \cal j}\left| \psi _ { l+1}-\psi _ { l}\right| ^{2}\right ) , \label{ldf}\end{aligned}\ ] ] where @xmath88 is the order parameter of the @xmath89th superconducting layer and the phenomenological constant @xmath90 is proportional to the josephson coupling between adjacent planes and @xmath91 .
the gauge @xmath92 is chosen in eq.(1 ) . in the vicinity of @xmath10 the ld functional
is reduced to the gl one with the effective mass @xmath93 along @xmath94-direction , where @xmath95 is the inter - layer spacing . in the gl region the fourth order term in ( 1 ) is omitted and the standard procedure@xmath96 to derive the fluctuation part of the free energy yields @xmath97 where the @xmath94-axis is along the z direction , @xmath98 and @xmath99 . by means of numerical derivation of eq.2 with respect to the field
one obtains the fluctuating magnetization @xmath83 vs. @xmath12 .
as shown in the inset of fig.4 , the magnetization curves in optimally doped ybco are satisfactorily fitted by @xmath83 derived in this way and evidence how the @xmath100 scenario of sf is obeyed on approaching @xmath10 , with a crossover from linear to non - linear field dependence occurring a few degrees above the transition .
correspondingly , the scaling arguments for @xmath24 anisotropic systems hold and @xmath101 vs. @xmath17 cross at @xmath102 , as already observed@xmath103 .
in contrast to optimally doped ybco , the magnetization curves for underdoped and overdoped compounds depart in a dramatic way from the ones expected on the basis of eq.2 .
in particular ( see fig.2a-3b-3c ) , even relatively far from @xmath10 , while for small fields ( h@xmath104t_c@xmath105linear in @xmath12 , upon increasing the field the magnetization shows an upturn and then @xmath106 decreases .
let us remind that in the gl weak fluctuation regime the saturation of the magnetization at high field has to be expected@xmath107 , the superconducting coherence being broken for fields larger than @xmath108 .
an estimate of the order of magnitude of the upturn field @xmath28 can be done from the analysis of the @xmath109 case@xmath96 , namely for superconducting granules of radius @xmath110 smaller than the coherence length @xmath16 . in this case
the order parameter is spatially homogeneous and the exact solution of the gl model can be found and yields @xmath111 it can be noticed that the most sizeable contribution to the magnetization comes from the fluctuations - induced sc droplets of radius of the order of @xmath112 , which imply most efficient screening.@xmath113 by assuming the condition of zero dimension for these droplets , from eq.3 with @xmath114 one derives an upturn field given by @xmath115 . for @xmath116 and @xmath117 in the range @xmath118 , @xmath28 is expected to be in the range of ten tesla . thus the magnetization curves in fig.s 3 can hardly be ascribed to the breakdown of the gl approach of the type commonly observed in bcs superconductors@xmath22 . in other words
, a description of fd based on the gl functional in principle should be suitable in ybco compounds for fields smaller than several tesla , particularly not too close to t@xmath71 , as in fact it is observed in optimally doped ybco.@xmath119 then one has to look for other explanations .
as already mentioned , a recent theory has been developed by sewer and beck@xmath51 at the aim to justify the unusual magnetization curves detected in underdoped ybco@xmath54 .
the theory assumes a frozen amplitude of the order parameter while phase fluctuations are taken into account . as a consequence
one has to deal both with thermally activated vortex loops and field induced vortex lines .
two major conclusions of general character can be outlined . for small field
the temperature dependence of the susceptibility is controlled by the vortex loops density @xmath121 , for which @xmath122 \eqnum{4}\ ] ] according to the xy model . for strong fields , instead
, the vortex line elements dominate , the vortex correlations between different layers become relevant and @xmath83 only slightly increases with h and finally it flattens . no upturn field is predicted , at least for @xmath123 . in fig.5 the data for @xmath124 , defined as @xmath125 ,
are shown to obey rather well to eq.4 , in correspondence to @xmath126 , in agreement with calculations yielding for the activation energy values around 10 @xmath10 ( see ref.@xmath51 and references therein ) .
@xmath127 turns out to depend only little from doping , being slightly field dependent .
it * * is necessary to * * mention that the temperature dependence of the susceptibility above @xmath128 differs from that * * one measured below @xmath10 , as shown in fig.5 .
however the magnetization curves , as the ones reported in figs . 3 and 4 , can not be accounted for by a theory which does not include an upturn with the field .
furthermore similar effects are found also in overdoped samples ( see figs.2a and 3a ) .
thus we are going to consider the second aspect possibly leading to an anomalous diamagnetism , the one related to charge inhomogeneities causing regions where the hole density is different from the average .
evidences of inhomogeneous structure of cuprates have been found by means of neutron and electron diffraction@xmath129 , as related to stripes and lattice effects or to local variation in the oxygen concentration , particularly near grain boundaries .
intrinsic inhomogeneities , with spatially dependent critical temperature have been considered as possible cause of pseudogap phenomena@xmath130 . a theory for high temperature superconductivity and of the pseudogap temperature dependence based on inhomogeneous charge distribution with site - dependent transition temperature has been recently formulated.@xmath131 in particular , ovchinnikov et al.@xmath50 have considered the anomalous diamagnetism above @xmath10 induced by non - uniform distribution of magnetic impurities , depressing @xmath10 but leaving islands which become superconductors above the resistive transition temperature .
an anomalous large diamagnetic moment results above @xmath10 and in this way the strong diamagnetic susceptibility observed in overdoped tl - based cuprates@xmath132 could be explained .
it should be stressed , however , that in this description@xmath133 the magnetization is linear in the field , since the condition of small field is implicitly assumed .
direct evidence of inhomogeneous magnetic domains showing diamagnetic activity above t@xmath71 has been obtained by iguchi et al.@xmath134 by scanning squid microscopy in underdoped lasco .
regions of few tens of @xmath135 , precursors of bulk superconductivity have been imaged in this remarkable work.@xmath134 in the light of the experimental findings and of the theoretical supports outlined above we consider now as a source of diamagnetism above @xmath10 the presence of locally superconducting droplets . from the volume susceptibility ,
let us say at @xmath136 ( see fig.1 ) , one deduces that a few percents of the total material being superconductor above the resistive transition , could actually justify the screening effects observed as fd .
a test of this hypothesis is obtained from the search of magnetic - history dependent effects .
it is known , in fact , that in ybco the irreversibility temperature is not far from @xmath10 and therefore if the anomalous fd has to be attributed to locally sc droplets then one should detect differences between zfc and field - cooled ( fc ) magnetization . in fig .
6 magnetization curves after zero field cooling and the correspondent values of @xmath83 obtained at the same temperature after cooling in the presence of a given magnetic field , are compared .
furthermore relaxation effects have been observed . in fig.7
it is shown how the negative magnetization depends on time , displaying a progressive decrease from the zfc value towards the one measured in fc condition .
the time constant for this relaxation process is close to the one measured in the critical state@xmath137 .
it can be remarked than in underdoped chain - disordered ybco ( fig.3c ) no upturn is observed and the zfc and fc magnetization curves almost coincide .
the explanation that will be supported from our theoretical picture is that the magnetization curves without hysteretic effects refer to superconducting droplets which are above the irreversibility temperature .
one could suspect that the occurrence of superconducting droplets results from trivial chemical inhomogeneities of the samples . as described in the section on experimental details many experimental checks allow us to rule out this hypothesis .
furthermore , samples grown with different procedures and already used by other authors , have been studied .
thus it is believed that the inhomogeneity does not mean the presence of macroscopic parts of the samples at different oxygen and/or calcium content , but it is rather intrinsic , as the ones evidenced in the experiments recalled above .
furthermore it should be remarked that the temperature dependence of the susceptibility above the bulk transition temperature is different from the one occurring in the superconducting state ( see fig.5 ) . in the following we are going to modify the theoretical description of sewer and beck@xmath51 , still keeping their basic idea of phase fluctuations but taking into account the presence of mesoscopic islands with non - zero average order parameter amplitude that can be below or above the local irreversibility temperature .
let us start , as in ref.@xmath51 , from eq.1 by evidencing the order parameter phase contribution @xmath138 = \frac{1}{s}{\sum_{l}}\int d^{2}r\left\ { { \em j}_{\shortparallel } \left ( { \bf \nabla } _ { \shortparallel } \theta -\frac{% 2ie}{c\hbar } { \bf a}_{\shortparallel } \right ) ^{2}+{\em j}_{{\em \perp } % } [ 1-\cos ( \theta _ { l+1}-\theta _ { l})]\right\ } \eqnum{5}\ ] ] where @xmath139 @xmath140 and @xmath141 are the order parameter phase coupling constants on the plane and between planes respectively . in this way the occurrence of superconducting droplets below the critical temperature is assumed , where the order parameter phase can fluctuate producing thermal excitations ( vortex and antivortex pairs in 2d , vortex loops in anisotropic model ) .
the potential vector @xmath142 in eq.5 describes both the magnetic field applied parallel to the @xmath94 - axis and the one induced by thermal fluctuations . by following the 2d coulomb gas theory , at each vortex is associated an effective charge @xmath143 and a vortex - antivortex pair has an energy @xmath144 @xmath145 , playing the role of an activation energy and thus yielding eq.4 . in order to refer to the anisotropic 3d model the vortex lines ( or the vertical elements of the vortex loops ) are correlated along the @xmath94 - axis for a length @xmath146 and a correction to @xmath147
was found selfconsistently . by considering , as usual , the partition function @xmath148 ) $ ] with @xmath149 , the susceptibility @xmath150 , where @xmath151 , is obtained as the sum of three contributions : @xmath152 where @xmath153 means the thermal average . in the gauge @xmath154 , @xmath155 being the c - axis direction ,
the homogeneous susceptibility is given by @xmath156 where @xmath157 .
\eqnum{8}\ ] ] in eq.(8 ) @xmath158 derives from the term @xmath159 of eq.(6 ) and it involves the current - current correlation function , as in ref.@xmath51 : @xmath160 \eqnum{9 } \\ & & \left\langle \left ( \nabla _ { x}\theta _ { l}(\rho ) -\frac{2\pi } { \phi _ { 0}}% a_{\shortparallel , x}\left ( { \bf r}\right ) \right ) \left ( \nabla _ { x}\theta _ { l^{\prime } } ( \rho ^{\prime } ) -\frac{2\pi } { \phi _ { 0}}a_{\shortparallel , x}\left ( { \bf r}^{\prime } \right ) \right ) \right\rangle \nonumber\end{aligned}\ ] ] with @xmath161 the number of layers and @xmath162 , @xmath163 being the average radius of the superconducting islands .
the x - component of the phase gradient is @xmath164 where @xmath165 and @xmath166 labels the position of each pancake
@xmath167 on the layer @xmath168 .
three terms are obtained by the evaluation of eq.(9 ) : @xmath169 , @xmath170 and @xmath171 ( the two terms due to the correlation between @xmath172 and @xmath173 give the same contribution being @xmath174 ) .
the first one involves the positional correlation function of the vortex line elements . in order to calculate it , sewer and beck@xmath51 introduced thestatic structure factor of a disordered vortex liquid . because of the weak interlayer coupling harmonic deviations of the vortex lines ( or loops ) along the @xmath155 direction are taken into account
this model can be used to describe also the vortex system in the glassy phase , below the irreversibility line temperature and therefore the same expression for @xmath175 is used here .
the evaluation of the term @xmath170 is straightforward and one has @xmath176 the further contribution @xmath171 , appearing due to the cross correlation between @xmath172 * * and * * @xmath177 * * and disregarded in ref*.*@xmath51 * , * can not be neglected below the vortex lattice melting temperature , where irreversibility effects occur . in this case one
obtains @xmath178 \right\rangle .\end{aligned}\ ] ] the thermal average is performed in the assumption that the vortices are uniformly distributed in the planes and the calculations are reported in appendix .
the expansion of @xmath171 in powers of @xmath179 gives @xmath180 the function @xmath181 in eq.8 , related to the third term of eq.6 , has been neglected in ref.@xmath51 .
it can be calculated as described in appendix , yielding @xmath182 .
\eqnum{12}\ ] ] it should be noted that the first term in @xmath183 , diverging for @xmath184 , exactly cancels out the @xmath185 term in the expansion of @xmath158 which appears from the structure factor . by using eqs .
( 10 ) , ( 11 ) , ( 12 ) and eq.(6 ) of ref.@xmath51 , from eq.(8 ) one finally obtains @xmath186 + \eqnum{13 } \\ & & \left [ -\frac{kt}{s\phi _ { 0}^{2}}\frac{1}{1 + 2n}\frac{(1+\delta \left ( \frac{h}{h^{\ast } } \right ) ^{2})^{2}}{n_{v}}-\frac{s^{2}\gamma ^{2}(1+n)}{% 1 + 2n}(1+\delta \left ( \frac{h}{h^{\ast } } \right ) ^{2})+\right .
\nonumber \\ & & \left .
\frac{47\pi r^{2}}{540}\frac{j_{\shortparallel } } { s}\left ( \frac{% 2\pi } { \phi _ { 0}}\right ) ^{2}\delta \left ( \frac{h}{h^{\ast } } \right ) ^{2}% \right ] q^{2 } \nonumber\end{aligned}\ ] ] with @xmath187 and @xmath188 is an effective critical field depending on the island size . to avoid unphysical divergences in the calculation of the susceptibility from eq.(7 ) , the first term in square brackets of eq.(13 ) has to be zero , giving a renormalization of @xmath147 due to both the anisotropy of the system and the presence of applied magnetic field : @xmath189 in view of the field - dependent vortex charge , the pair energy ( in the limit @xmath190 ) becomes @xmath191 . according to eq.(4 ) the thermally - excited vortex pair density turns out field dependent .
this field dependence , formally derived in our description , is significantly different from the one assumed in ref.@xmath51 .
finally the diamagnetic susceptibility is obtained in the form @xmath192 in the limit @xmath193 a good agreement of the susceptibility and its temperature dependence with the experimental findings is again achieved .
the main differences between our susceptibility in eq.(15 ) and the one given in ref.@xmath51 consists in the presence of the factor @xmath194 and of the third , positive term . this term can give an inversion in the sign of the susceptibility corresponding to an upturn in the magnetization curves .
this phenomenon depends on the dimension of the islands and @xmath195 ( i.e. the occurrence of the upturn ) requires @xmath196 where @xmath197 depends on some characteristics of the material . by choosing @xmath198 , the interlayer distance @xmath199 , @xmath200 , @xmath201 , which are typical values for ybco , for @xmath202 one estimates @xmath203 . in this case
the solutions of the equation @xmath195 is @xmath204 and by considering the experimental value @xmath205 g , the effective critical field turns out @xmath206 .
the isothermal curves can be obtained from eq.(15 ) by means of numerical integration .
the shape of the magnetization curve depends on the parameters in the susceptibility and by using the values quoted before , with @xmath207 one derives the behaviour sketched in figs.3a for an island below the irreversibility line . the same parameters with @xmath208 , @xmath209 and @xmath210 lead to the curve shown in fig.3c for the magnetization of the island above irreversibility .
finally we discuss the differences observed in the magnetization curves between chain - ordered and chain - disordered ybco compounds and the relevant observation by carballeira et al.@xmath42 of magnetization curves -m@xmath65 vs h in underdoped lasco ( sr content @xmath5=0.1 , t@xmath71(0)=27.1 k ) with no upturn field . a role of the chains on favouring the nucleation of local superconducting droplets above t@xmath71
is conceivable .
in fact , in chain - ordered compounds the droplets appear to have the irreversibility temperature higher than the ones in chain - disordered compounds , as it is evidenced by the difference in the magnetization curves ( see figs .
3a - b - c ) .
we remind that the inversion in the sign of the susceptibility is related to the third term in eq.15 and thus to the term @xmath171 .
the amount of impurities and/or imperfections acting as nucleation centers might also play a role .
furthermore the degree of under or over - doping is also involved since a marked variation of t@xmath71 with n@xmath211 is evidently crucial .
it is noted that in the measurements by carballeira et al.@xmath42 in lasco at t@xmath71=27.1 k the magnetization curves show only a weak tendency to saturation while in lasco at t@xmath21218 k ( therefore more underdoped ) scanning squid microscopy by iguchi et al@xmath134 does evidence diamagnetic effects to associate to locally superconducting droplets .
these droplets should imply a contribution to the magnetization curves similar to the one detected by us in ybco compounds . future research work will have to explore these interesting aspects and the differences until now present between lasco and ybco .
by means of squid measurements in @xmath59 family a non conventional fluctuating diamagnetism has been observed in overdoped and in underdoped compounds .
compared to optimally doped ybco , a large enhancement of the diamagnetic susceptibility occurs and no anisotropic gl functional or scaling arguments can justify the isothermal magnetization curves .
the recent theory@xmath51 for phase fluctuations of the order parameter in a layered liquid of vortices has been revised and it appears to justify some aspects of the anomalous fd in non - optimally doped ybco , particularly the precritical temperature activated behaviour of the susceptibility in the limit of zero field .
other experimental observations , noticeably the upturn in the field dependence of the isothermal fluctuating magnetization and history - dependent effects , indicate the role of mesoscopic charge inhomogeneities in inducing local , non - percolating , superconducting droplets . on the basis of both types of experimental findings
we have extended the theory of phase fluctuations in the presence of non - zero order parameter .
the terms leading to a novel and relevant dependence of the fluctuating magnetization from the magnetic field were included in the scheme .
the field - related corrections are different when the superconducting droplets are below or above the local irreversibility temperature . in this way
most of the experimental findings have been justified .
thanks are due to f.borsa , f. licci and s. kramer for having provided some of the samples used in the present work and for useful discussions .
f. cordero is thanked for the deoxygenation of the underdoped samples and s.sanna for the resistivity measurements as a function of temperature .
one of the authors ( a.v . )
thanks h.beck and a.sewer for interesting discussions during his visit to neuchatel university .
d.e . has carried out the work in the framework of the stage program of the infm national school ( 2000 ) .
@xmath14 see fluctuation phenomena in high temperature superconductors , edited by m.ausloos and a.a.varlamov , kluwer academic publishers , netherlands , 1997 ; see also a.a.varlamov , g.balestrino , e.milani and d.v.livanov , adv .
48 , 655 ( 1999 ) .
@xmath213m.tinkham introduction to superconductivity mc graw - hill , new york 1996 , chapter 8 .
@xmath25 j.kurkijarvi,v.ambegaokar and g.eilenberger , phys.rev b 5 , 868 ( 1972 ) .
@xmath214 see a.i .
larkin and a.a.varlamov , fluctuation phenomena in superconductors in physics of conventional and non - conventional superconductors eds .
k .- h.bennemann and j.b .
ketterson , springer verlag ( 2001 ) .
@xmath215 l.n .
bulaevskii , int .
j. of modern phys .
b 4 , 1849 ( 1990 ) .
@xmath216 l.n .
bulaevskii , m.ledvig and v.g .
kogan , phys .
68 , 3773 ( 1993 ) .
@xmath217 v.g .
kogan , m. ledvij , a. yu .
simonov , j. h. cho , and d. c. johnston , phys .
70 , 1870 ( 1993 ) .
@xmath218 m.a .
hubbard , m.b . salamon and b.w .
veal , physica c 259 , 309 ( 1995 ) .
@xmath219 c.baraduc , a. buzdin , j.y .
henry , j.p . brison and l. puech , physica c 248 , 138 ( 1995 ) .
@xmath220 a.e.koshelev , phys .
rev b 50 , 506 ( 1994 ) .
@xmath221 a.budzin and v.dorin , in fluctuation phenomena in high temperature superconductors ,
ref.1 , p.335 .
a detailed theoretical analisys of the field dependence of the gaussian fluctuations in layered superconductors has been given by t.mishonov and e. penev , intern .
j.of modern phys .
b 14 , 3831 ( 2000 ) .
@xmath34 t. schneider and u. keller , inter .
j. of modern phys .
b 8 , 487 ( 1993 ) and references therein ; t. schneider and u. keller , physica c 207 , 336 ( 1993 ) ; see also t.schneider and j.m .
singer , in phase transitions approach to high temperature superconductivity , imperial college press 2000 , chapter 6 .
@xmath103 a.junod , j - y genoud , g.triscone and t.schneider , physica c 294 , 115 ( 1998 ) .
@xmath222 s.salem-sugui and e.z .
dasilva , physica c 235 , 1919 ( 1994 ) .
@xmath36 k.kanoda , t. kawagoe , m. hasumi , t. takahashi , s. kagoshima and t. mizoguchi , j.phys .
japan 57 , 1554 ( 1988 ) .
@xmath223 p.carretta , a.lascialfari , a.rigamonti , a.rosso and a.a.varlamov , inter
. j. of modern phys .
b 13 , 1123 ( 1999 ) .
@xmath224 d.babic,j.r.cooper , j.w .
hodby and chen changkang , phys .
b 60 , 698 ( 1999 ) .
@xmath54 p.carretta , a.lascialfari , a.rigamonti , a.rosso and a.a.varlamov , phys .
rev.b 61 , 12420 ( 2000 ) .
@xmath42 c.carballeira , j. mosqueria , a.revcolevschi and f.vidal , phys .
84 , 3157 ( 2000 ) .
@xmath44 b.rosenstein , b.y .
shapiro , r. prozorov , a. shaulov , y. yeshurun , phys.rev .
b 63 , 134501 ( 2001 ) . @xmath225 a.rigamonti and p. tedesco , lecture notes at the infm national school in structure of matter ( villa gualino , torino , september 2000 ) and scientifica acta , university of pavia , xv , 49 ( 2000 ) .
@xmath50 yu.n .
ovchinnikov , s.a.wolf and v.z .
kresin , phys.rev .
b 60 , 4329 ( 1999 ) . @xmath51 a.sewer and h.beck , phys .
rev.b 64 , 014510 ( 2001 ) .
@xmath53 p.carretta , a lascialfari , a.rigamonti , f.teodoli , f.bolzoni e f.licci , inter . j.of modern phys .
b 14 , 2815 ( 2000 ) .
@xmath226 u.welp , s. fleshler , w. k. kwok , r. a. klemm , v. m. vinokur , j. downey , b. veal , and g. w. crabtree , phys.rev .
67 , 3180 ( 1991 ) .
@xmath227 c. torron , a. diaz , a. pomar , j.a .
veira and f. vidal , phys . rev .
b49 , 13143 ( 1994 ) @xmath228 j. mosqueira , c. carballeira , m.v .
ramallo , c. torron , j. veira and f. vidal , europhys .
53 , 632 ( 2001 ) @xmath74 j.l .
tallon , c. bernhard h. shaked r. l. hitterman and j. d. jorgensen , phys .
b 51 , 12911 ( 1995 ) .
@xmath113 j.p.gollub , m. r. beasley , r. callarotti and m. tinkham , phys .
b 7 , 3039 ( 1973 ) .
@xmath229 m.gutmann , s.j.l .
billinge , e.l .
brosha and g.h .
kwei , phys . rev .
b 61 , 11762 ( 2000 ) .
@xmath230 e.bozin .
g.kwei , h.takagi and s.billinge , phys.rev .
84 , 5856 ( 2000 ) .
@xmath231 z.akase , y.tomokiyo , y.tanaka and m.watanabe , physica c 339 , 1 ( 2000 ) .
@xmath232 yu.n .
ovchinnikov , s.a.wolf and v.z .
kresin , phys.rev .
b 63 , 064524 ( 2001 ) .
@xmath233 e. v. l. de mello , e.s .
caixeiro and j.l .
gonzales , cond - mat 0110519 ( 24 oct 2001 ) ; ibidem , cond - mat 0110479 ( 22 oct 2001 ) @xmath132 c.bergemann , a. w. tyler , a. p. mackenzie , j. r. cooper , s. r. julian and d. e. farrell , phys .
b 57 , 14387 ( 1998 ) .
@xmath134 i. iguchi , t. yamaguchi and a. sugimoto , nature 412 , 420 ( 2001 ) @xmath137y.yeshurun , a.p .
malozemoff , a. shaulov , rev . of modern phys .
68 , 911 ( 1996 ) .
to derive eqs.(11 ) and ( 12 ) the following thermal average must be calculated : @xmath234 indicating with @xmath235 ( @xmath236 ) the number of vortex line elements parallel ( antiparallel ) to the field the first term gives @xmath237 the sum in the second term can be split in two parts which separately count the vortex line elements parallel and antiparallel to the field : @xmath238 one can assume that the vortices are uniformely distributed in the planes and that the @xmath239 components of their positions are distributed on a line , separated each other by a distance @xmath240.then , the @xmath241th vortex is in the mean position @xmath242 , with @xmath243 .
@xmath244 by considering @xmath245 * * one finally finds @xmath246 then ( a1 ) can be written @xmath247
fig.1 some magnetization data in low field , parallel to the c - axis , as a function of temperature in oriented powders of @xmath248 .
the values of the magnetization measured from 250 k down to 90 k , with a positive pauli - like temperature independent term , are not reported in the figure . in the insets the blow - up for the estimate of @xmath40
is shown .
fig.2 constant field ( @xmath249 ) magnetization vs. temperature in overdoped ( @xmath250 and @xmath251 ) ( a ) and in underdoped ( @xmath252 and @xmath253 ) ( b ) ybco compounds . for comparison in the part
a ) of the figure the behaviour of @xmath83 in optimally doped ybco is shown .
fig.3 isothermal diamagnetic magnetization vs. @xmath12 , after zero - field cooling ( zfc ) to a certain temperature above @xmath254 a ) sample at @xmath250 , @xmath255 ( @xmath256 ) ; b ) sample at @xmath257 , @xmath258 ( @xmath259 ) ; c ) chain disordered underdoped ybco at @xmath253 and @xmath252 ( @xmath260 ) for zfc ( circle ) and field - cooled ( fc ) ( up - triangle ) conditions .
the solid lines in part a ) of the figure correspond to the diamagnetic susceptibility estimated in the limit of zero field .
the solid lines in figs.3a and 3c are the theoretical behaviours according the mechanisms described in the text for droplets below and above the local irreversibility temperature .
fig.4 comparison of the magnetization curves @xmath83 vs. @xmath12 ( after zfc ) in overdoped ybco : ca ( @xmath250 , @xmath251 ) with the ones in optimally doped ybco ( inset ) , for similar reduced temperature @xmath117 .
the solid lines fitting the data in optimally doped ybco are derived from eq.2 in the text by means of numerical derivation , correspond to anisotropic parameter r=0.1 and evidence the 3d linear and 3d non - linear regimes .
fig.5 susceptibility as a function of the inverse temperature showing the activated temperature behavior in the sample at @xmath251 and @xmath250 .
analogous temperature dependence has been observed in underdoped ybco compounds .
the dashed lines are obtained by transferring above the superconducting temperature @xmath40 the temperature behavior of the bulk susceptibility measured below @xmath10 and by normalizing the data at @xmath261 .
fig.6 magnetization curves in ybco : ca at @xmath250 and @xmath251 by cooling in zero field to a given temperature [ a ) @xmath202 ; b ) @xmath262 above the resistive transition temperature and then applying the field ( zfc ) and after the application of the field at room temperature , cooling at the same temperature and measuring the correspondent magnetization ( fc ) .
the volume susceptibility in the limit h@xmath2630 is reported .
relaxation of the raw magnetization after zfc and then sudden application of a magnetic field of @xmath264 g , in ybco : ca at @xmath250 and @xmath265 , at @xmath202 a ) short - term relaxation ; b ) long - term relaxation . from the comparison of the zfc and fc magnetization in h=20 g ( see inset ) an irreversibility temperature of the locally superconducting droplets at highest @xmath10 can be estimated around 90 k. in part a ) of the figure the solid line is the sketchy behavior of the relaxation of the magnetization detected by yeshurun et al .
( ref.@xmath232 ) in the critical state , in optimally doped ybco .
table i : superconducting transition temperature in overdoped and underdoped y@xmath0ca@xmath1ba@xmath52cu@xmath3o@xmath4 and estimated number of holes per cuo@xmath52 unit | squid magnetization measurements in oriented powders of y@xmath0ca@xmath1ba@xmath2cu@xmath3o@xmath4 , with @xmath5 ranging from @xmath6 to @xmath7 , for @xmath8 and @xmath9 , have been performed in order to study the doping dependence of the fluctuating diamagnetism above the superconducting transition temperature @xmath10 . while for optimally doped compounds the diamagnetic susceptibility and the magnetization curves @xmath11 ) vs. @xmath12 are rather well justified on the basis of an anisotropic ginzburg - landau ( gl ) functional , in underdoped and overdoped regimes an anomalous diamagnetism is observed , with a large enhancement with respect to the gl scenario .
furthermore the shape of magnetization curves differs strongly from the one derived in that scheme .
the anomalies are discussed in terms of phase fluctuations of the order parameter in a layered system of vortices and in the assumption of charge inhomogeneities inducing local , non percolating , superconducting regions with @xmath13 higher than the resistive transition temperature @xmath10 .
the susceptibility displays activated temperature behavior , a mark characteristic of the vortex - antivortex description , while history dependent magnetization , with relaxation after zero - field cooling , is consistent with the hypothesis of superconducting droplets in the normal state .
thus the theoretical picture consistently accounts for most experimental findings .
[ theorem]acknowledgement [ theorem]algorithm [ theorem]axiom [ theorem]claim [ theorem]conclusion [ theorem]condition [ theorem]conjecture [ theorem]corollary [ theorem]criterion [ theorem]definition [ theorem]example [ theorem]exercise [ theorem]lemma [ theorem]notation [ theorem]problem [ theorem]proposition [ theorem]remark [ theorem]solution [ theorem]summary pacs : 74.40.+k , 74.70.vy,74.25.ha |
to quantitatively describe an evolving ecosystem requires some general principle that applies equally to all species .
a candidate for such a principle is to consider the relationship between an organism s genotype and its _ fitness _ , which is some measure of the expected number of genes passed back into the species gene pool ( dawkins 1983 ) .
each point in the multidimensional space of all possible genotypes can be assigned its own fitness value , forming a _ fitness landscape _ , as schematised in fig .
[ f : fit_land ] .
the process of evolution can then be described as a walk over this landscape in the direction of increasing fitness . rather than try to calculate the fitness for each genotype from first principles , clearly a hopeless task , kauffman has assumed that the relationship is so complex that it can be well approximated by random variables ( kauffman 1993 ) .
this leads to the concept of _ rugged fitness landscapes _ , where `` rugged '' refers to the large variations in fitness that can result from small changes in genotype .
models based on the bak - sneppen approach assume that each species moves on its own rugged fitness landscape , eventually becoming trapped in the region of a local maximum .
if the landscape is fixed , then the species can only evolve by moving to a different maximum .
suppose that a species labelled @xmath5 is at a local maximum which is separated from nearby maxima by fitness barriers of heights @xmath6 , where and the @xmath7 are ordered such that . over time , fluctuations in the species fitness may bring it into the vicinity of one of its barriers , allowing it to crossover to a different maximum .
this constitutes an evolutionary event in which the species changes from one typical genotype to another . for uncorrelated fluctuations ,
the expected time @xmath8 to crossover a barrier of height @xmath7 will be given by an arrhenius equation of the form @xmath9 where the constant @xmath10 fixes the timescale and is analogous to temperature .
since the landscape is rugged , a species that escapes from one maximum will soon become trapped by another , and will find itself surrounded by an entirely new set of barriers @xmath7 . in the limit , ( [ e : se_tevolve ] ) implies that it will always be the species @xmath5 with the smallest @xmath7 in the system that evolves first , and that the other species will have moved no appreciable distance towards their own barriers by the time this occurs .
thus the ecosystem will consist of species that infrequently hop between maxima at a rate that depends upon the @xmath6 , but are otherwise essentially static .
the evolution of species @xmath5 will alter the landscapes of all those species @xmath11 linked to it in the ecosystem , for instance via predator - prey or host - parasite relationships . although each species @xmath11 will in general have to move on their new landscapes before finding a new maximum , it is a further approximation of the theory that this can be ignored and only the barriers @xmath12 are affected .
the original bak - sneppen model is defined purely in terms of the smallest barriers @xmath13 , which are arranged on a lattice in such a way that interacting species occupy adjacent lattice sites . the evolutionary process described in the previous paragraphs then reduces to the dynamical interaction between adjacent @xmath14 .
the system is static until the site with the smallest @xmath14 evolves , when @xmath14 and all of the @xmath15 in adjacent sites @xmath11 are assigned new values .
the system is again static until another site evolves , and so on .
it has been shown that the essential system behaviour is insensitive to details such as the choice of probability distribution for the @xmath14 ( bak 1993 , paczuski 1996 ) .
this _ robustness
_ relates to the universality of the critical state , and is essential if such a simple model is to faithfully describe real ecosystems .
the bak - sneppen model can be enhanced by a more detailed consideration of the fitness landscapes and their interaction .
three features absent from the original model will be considered here , namely _ speciation _ , _ extinction _ and _ external noise_. each feature is described in general terms below before the new model is fully specified . _ speciation : _ speciation occurs when two sub - populations reach a state of reproductive isolation and should be considered as separate species ( maynard - smith 1993 , ridley 1993 ) .
for instance , two groups that are reunited after prolonged geographical isolation may have evolved so much in different directions that they are unable to produce viable offspring . up until now
, a species has been described as simply occupying a region of genotype space .
more detailed analysis shows that a population forms a `` cloud '' of points of roughly equal fitness around the local maximum ( kauffman 1993 ) .
normally the whole population crosses over the same barrier @xmath14 , but if then it is possible that part of the population will instead cross over the barrier @xmath16 . if this happens , the two subpopulations will move to different maxima and a speciation event will have occurred . a simple criterion for speciation is to say that species @xmath5 will branch into two subspecies if @xmath17 when it evolves to a new form , where @xmath18 is a constant parameter .
further barriers could also be considered to incorporate the simultaneous splitting into three or more subspecies , but such events will be very rare and are ignored here . _ extinction : _ the system size would increase without limit if only speciation were allowed , so some mechanism is required by which a species can be made extinct and permanently removed from the system .
the original bak - sneppen model does not distinguish between this form of extinction and _ pseudo - extinction _ , which is where a rapidly evolving species disappears from the fossil record if its intermediate forms are not recorded .
what is required is some criterion for true extinction defined purely in terms of the individual species fitness landscapes , analogous to ( [ e : se_spec ] ) .
it is not clear how this may be achieved . instead
, a heuristic approach is adopted here , which is to say that a species @xmath11 is made extinct if it is linked to the species with the minimum barrier , and has @xmath15 greater than some threshold value .
this proves to be the simplest choice for which the system size does not diverge . _
external noise : _ a fitness landscape is ultimately a function of the species itself , the species with which it interacts , and any factors external to the ecosystem , so fluctuations in the inorganic environment can also cause the fitness barriers to change .
examples include local disturbances such as volcanic eruptions or the formation of a new river , to global events including meteor impacts and changes in the sea level .
the interactions between species have already been incorporated into the model , but no allowance has yet been made for these abiotic factors .
continuing with the philosophy that changes in fitness can be approximated by random variables , external influences are assumed to alter the fitness landscapes by an amount @xmath19 per unit time , where @xmath20 is a new parameter .
more precisely , every species in the system will have their barriers altered by an amount @xmath21 where the @xmath22 are uniformly distributed on @xmath23 $ ] and uncorrelated in time .
external effects will occur on a separate timescale to the evolutionary processes in ( [ e : se_tevolve ] ) , but for simplicity both timescales are fixed at the same constant rate in this model . it remains to be decided how interacting species are linked together .
the original bak - sneppen model placed the species on a regular crystalline lattice in which interacting species occupy adjacent sites , but this is not flexible enough to incorporate the addition of new species to the system and is of no use here .
real food webs have a much more involved structure , and if the full range of interactions is allowed rather than just links in the food chain , then it appears that a great many species interact at least weakly ( hall 1993 , caldarelli 1998 ) . trying to model this
would only serve to draw attention away from the main motivation for the new model , which is to allow a variable system size .
instead , we adopt the mean field approach in which each species @xmath5 interacts with @xmath24 other species @xmath11 chosen at random from the system .
the species @xmath11 are reselected at every time step . the extended model can now be fully specified .
the ecosystem consists of @xmath25 species labelled by .
each species occupies the region around a local maximum on a rugged fitness landscape , and is separated from nearby maxima by barriers of various heights @xmath7 .
the larger barriers can be ignored since they will rarely contribute to the dynamics , but at least two must be retained if speciation is to be incorporated .
hence each species is defined by its two smallest barriers @xmath14 and @xmath16 , which are uniformly distributed over the range [ 0,1 ] and then ordered so that .
the following steps ( i)(vi ) are iterated for every time step .
\(i ) _ evolution : _ the smallest @xmath14 in the system is found and marked for evolution .
it will move to a new maximum in step ( vi ) .
\(ii ) _ speciation : _ if the single species marked for evolution has @xmath26 , then a new species is introduced to the system with random barriers . @xmath27 .
\(iii ) _ interaction : _ @xmath24 other species are chosen at random from the remaining @xmath28 in the system .
they will be assigned new barriers in step ( vi ) .
\(iv ) _ extinction : _ if any the of the @xmath24 interacting species has @xmath29 , it is removed from the system . @xmath30 .
\(v ) _ external noise : _ every barrier @xmath7 in the system is transformed according to ( [ e : se_noise ] ) , and reordered if necessary .
\(vi ) _ new barriers : _ the species marked for evolution in step ( i ) and the @xmath24 interacting species from step ( iii ) are assigned new random barriers , ordered such that . with such specific definitions of general processes ,
it is obviously important to check that the model is robust to any arbitrary choices . to test this ,
the simulations have been repeated with various changes to the rules , and in no case was any qualitative deviation observed .
for instance , different values for the extinction threshold in step ( ii ) give the same behaviour , even if the threshold value varies in time around a fixed mean .
both @xmath18 and @xmath20 were chosen from uniform , gaussian and exponential distributions , again with no apparent change in behaviour .
since the model appears to be robust , further discussion will be restricted to the simple set of rules given in . the threshold value for extinction
was fixed at 1 to minimise the number of new parameters .
before continuing , it should be pointed out that the algorithm presented in steps above is not exactly the same as that described in our previous exposition of this work ( head 1997 ) .
this earlier model assumed that all @xmath31 species `` mutate '' ( evolve ) at every time step , whereas it is of course just the species with the minimum barrier that evolves .
the corrected model studied here behaves in much the same way as its previous incarnation , except that the number of species is now only weakly dependent on the connectivity @xmath31 .
this is in agreement with data for real multispecies communities , as discussed further in the next section .
the quantity of interest is the total system size @xmath25 .
this varies in a manner that depends upon the choice of values for the parameters @xmath31 , @xmath18 and @xmath20 , as described below .
@xmath32 : steps ( ii ) , ( iv ) and ( v ) never feature in the time evolution of the system and the larger barriers @xmath16 are redundant .
the @xmath14 interact in the same way as the original bak - sneppen model , the only difference being that each @xmath14 is the smaller of two uniform distributions on [ 0,1 ] and so is distributed according to , . since the model is robust to the choice of probability distribution , this difference is not important .
@xmath25 remains fixed at its initial value .
@xmath32 : there are no interactions , and the species that evolves will always have so extinction is impossible . if or remains fixed if .
@xmath32 _ @xmath33 , and _ : there is speciation but no extinction , @xmath34 .
@xmath32 _ @xmath33 , and _ : there is extinction but no speciation , @xmath35 .
@xmath32 _ @xmath33 , @xmath36 and @xmath37 _ : @xmath25 fluctuates around some constant value @xmath0 which is independent of the initial system size . an example is given in fig .
[ f : se_n_infty ] .
note that if , @xmath0 is so small that statistical fluctuations will eventually send and the simulation is over .
that @xmath0 should exist at all is by no means obvious , since @xmath25 does not explicitly appear in the rules for speciation and extinction .
it exists because of the external noise of order @xmath20 , which is just as likely to push two barriers apart as to bring them together and so does not affect the rate of speciation . however , the noise acts asymmetrically on barriers near the threshold for extinction , tending to push species over this threshold into the small tail corresponding to those species that will be made extinct when next selected .
since every species is subjected to external noise at every time step , the rate of extinction increases with @xmath38 whilst the rate of speciation remains roughly constant .
a steady state will be found when these two rates balance .
this qualitative reasoning is confirmed by the analysis in sec .
[ s : se_anal ] .
the parameters @xmath18 and @xmath20 are abstract quantities defined purely in terms of the model , so it is not possible to estimate their values for real ecosystems .
nonetheless it is intuitively reasonable to assume that speciation and extinction events are rare , and therefore both @xmath18 and @xmath20 should be small .
the number of links per species @xmath24 has been measured for real communities , and according to some studies is independent of the system size ( hall 1993 , kauffman 1993 ) .
this is in agreement with numerical simulations of the model , which shows only a weak dependence on @xmath31 from the range to , as given in table [ t : se_num ] ( at end of document ) .
there is a small peak around , which also corresponds to the most common value of @xmath31 observed in nature .
however , the data for real ecosystems is based on food webs whereas the bak - sneppen approach considers all direct interactions between species , so it is not clear how far this comparison can be taken . turning to consider global ecosystems , the fossil record for all marine organisms highlights the possibility of a statistical steady state throughout much of the palaeozoic era
. the total number of ( families of ) species fluctuates around a roughly constant value up until the mass extinction at the end of the permian period , after which the system size increased beyond its earlier levels and is still increasing today ( benton 1995 ) .
it could be conjectured that the new species that emerged after the end - permian extinction were on average either more likely to speciate , or less susceptible to external noise , or both , which should result in an increased system size according to the model .
the data for continental organisms is less clear and if anything shows a continuing increase in diversity at varying rates .
the distribution of the magnitude of the change in @xmath38 per unit time is poisson to first order , implying that the speciation and extinction events are uncorrelated for .
however , the distribution is not _ precisely _ poisson , which is presumably due to the tendency for @xmath38 to drift towards @xmath0 when is large .
an example is presented in fig .
[ f : se_jump_size ] . that the distribution was not power law is disappointing , but perhaps unsurprising given that the interactions between species are randomised at every time step , making it difficult for the system to self - organise .
it is possible that a spatially extended model might allow for correlations to build up towards a critical state and a power law to be recovered , but this must remain as speculation at present .
it is possible to derive the dependence of @xmath0 on the parameters @xmath31 , @xmath18 and @xmath20 by extending the mean field solution of the original bak - sneppen model ( flyvberg 1993 ) . in theory
this approach could give the exact solution , since the interacting species are selected at random in the new model and so it is mean field by definition .
however , the increased dynamical complexity means that only the first order parameter dependence has be calculated .
* standard model with two barriers per species * the original solution was based on a single barrier per species . before tackling speciation and extinction ,
it must first be shown how the mean field approach can be modified to handle pairs of barriers .
define @xmath39 to be the probability that a randomly selected species has one barrier in the range [ x , x+dx ) and the other in [ y , y+dy ) .
note that this refers to the barriers @xmath7 _ before _ ordering , so @xmath40 can be less that or greater than @xmath41 .
the probability for a species to have _ both _ barriers greater than is represented by @xmath42 , which is related to @xmath43 by @xmath44 since @xmath45 for values of @xmath40 or @xmath41 outside the range [ 0,1 ] , for and for
. the species with the smallest barrier can be any of the @xmath38 in the system , as long as all of the remaining species have larger barriers .
hence @xmath46 , the probability distribution for the species with the smallest barrier , is given by @xmath47 at each time step , @xmath43 will change by an amount @xmath48 which is given by the master equation @xmath49 the first term on the right hand side of ( [ e : se_evolution ] ) accounts for the evolution of the species with the lowest barrier , the second for the new barriers assigned to the @xmath24 species with which it interacts , and the third term handles the @xmath31 new pairs of barriers . in the statistical steady state ,
@xmath50 and , using ( [ e : se_pmin ] ) and ( [ e : se_evolution ] ) , @xmath51 the solution to ( [ e : se_steady ] ) depends upon the behaviour of @xmath52 as ( flyvberg 1993 ) .
if @xmath53 , then the term proportional to @xmath52 vanishes and @xmath54 conversely , if either @xmath40 or @xmath41 is so small that @xmath55 , then the second term in ( [ e : se_steady ] ) will be @xmath56 and , so @xmath57 and hence from ( [ e : se_pmin ] ) , @xmath58 each solution applies in different regions of the @xmath59 plane , which , for large @xmath38 , will be separated by sharply defined boundaries .
these boundaries can be found by remembering that both @xmath60 and @xmath61 are probability distributions and normalise to one .
to first order in @xmath62 , the full solutions are @xmath63 hence the species with the smallest barrier will always be found in the region where , and its interacting species will always come from the region corresponding to . *
analysis for @xmath18 and @xmath20 non - zero but small * when either or , the system size @xmath38 becomes a function of time and the algebra quickly becomes prohibitive .
the simpler and more intuitive approach adopted here is to initially ignore speciation and extinction altogether and only incorporate the external noise of order @xmath20 .
this leads to new solutions for @xmath43 and @xmath46 , from which the rates of speciation and extinction can be calculated even though they are no longer dynamically involved .
the natural system size @xmath0 is then the value of @xmath38 for which the two rates balance .
the analysis presented below assumes that @xmath20 is small ; large values of @xmath20 and @xmath18 are considered at the end of this section .
the effect of the external noise will be to perturb the solutions for @xmath60 and @xmath61 given in ( [ e : se_solp ] ) and ( [ e : se_solpmin ] ) , as schematised in fig .
[ f : se_mfsol ] .
the master equation ( [ e : se_evolution ] ) must be modified in two ways .
first , the external noise can cause barriers to move outside of the range [ 0,1 ] , so the range of possible @xmath40 and @xmath41 must be extended .
however , the barriers are still assigned values in the range [ 0,1 ] and the term for new barriers must be altered accordingly .
secondly , a term for the noise itself must be included .
the new steady state equation is @xmath64 @xmath65 the theta functions in the first term ensure that new barriers lie in the range [ 0,1 ] .
the last term on the right hand side of ( [ e : se_stdynoise ] ) accounts for the external noise , where @xmath66 is the laplacian operator .
a full derivation of this term is given in the appendix .
_ rate of extinction : _ each of the @xmath24 random neighbours will be made extinct if it has @xmath67 and @xmath68 .
thus the rate of extinction @xmath69 is given by @xmath70 where the integral is over the region @xmath71 in fig .
[ f : se_mfsol ] .
strictly speaking , the distribution in this equation should be @xmath72 , but this distinction can be ignored for large @xmath38 . when both barriers are large , @xmath73 and ( [ e : se_stdynoise ] ) can be simplified by the transformation @xmath74 to give @xmath75 for either @xmath76 or @xmath77 negative , corresponding to @xmath67 or @xmath68 , the second term on the right - hand side of ( [ e : se_newvar ] ) vanishes and the equation can be solved by separation of variables .
coupled with the boundary conditions @xmath78 for @xmath79 or @xmath80 , the solution is @xmath81 where @xmath82 is an arbitrary constant .
whatever the value of @xmath82 , it must be independent of @xmath31 and @xmath20 since these parameters do not appear in ( [ e : se_newvar ] ) .
transforming back into the original variables gives the explicit parameter dependence , @xmath83 substituting this into ( [ e : se_ke ] ) gives @xmath84 _ rate of speciation : _ it has not been possible to find a solution to ( [ e : se_newvar ] ) for and .
the variable separable solution does not behave correctly , and other methods tried have been fruitless .
instead , the @xmath85 solution will be used as a first approximation .
the rate of speciation @xmath86 will be proportional to the density of species with @xmath87 .
since only the species with the minimum barrier can speciate , @xmath88 substituting the explicit expression for @xmath46 from ( [ e : se_solpmin ] ) into ( [ e : se_k_s ] ) gives @xmath89 for small @xmath18 . with @xmath37
, @xmath46 broadens and so the real rate of speciation will decrease for larger @xmath20 .
the value of @xmath0 can now be found up to parameter dependence .
the rates of speciation and extinction balance when , and therefore @xmath90 this implies that should be roughly constant .
this quantity has been calculated from the numerical simulations and is shown in table [ s : se_num ] .
the agreement is good for variations in @xmath20 , but less so for @xmath31 and @xmath18 .
this most probably reflects the first order approximation used in deriving @xmath86 in ( [ e : se_ks ] ) .
* either @xmath18 or @xmath20 large * for the sake of completeness , the equivalent expression to ( [ e : se_ninfty ] ) will now be derived for large @xmath18 or @xmath20 , although such values bear no relevance to actual systems .
if @xmath18 is large but @xmath20 small , the system size rapidly increases and with it the expected time a species will move under the influence of external noise before being assigned new barriers . similarly ,
if both @xmath18 and @xmath20 are large , then the system behaviour is also dominated by the external noise .
this is called the _ noise dominated regime_. if @xmath18 is small but @xmath20 large , @xmath0 becomes so low that fluctuations will eventually drive every species in the system to extinction . in the noise dominated regime , @xmath43 will no longer be just a small perturbation around the original solution but will extend to large positive and negative values in both the @xmath40 and @xmath41 directions .
since the external noise is isotropic , @xmath43 will be symmetric about the @xmath40 and @xmath41 axes and at most 1 in 4 species have both barriers in the @xmath91 region .
hence the rate of extinction will approach its upper bound value of @xmath92 when a barrier is assigned a new value in the range [ 0,1 ] , it undergoes an unbiased random walk until it is again assigned a new value and brought back to near the origin .
the average number of steps in this walk will be and , since the average step size is @xmath19 , an analogy with a one - dimensional random walker implies that the total distance travelled will be @xmath93 ( papoulis 1991 ) .
this gives the width of the barrier distribution in both the @xmath40 and @xmath41 directions .
the number of species in the infinite strip @xmath87 is inversely proportional to the width of @xmath43 , so the rate of speciation is now given by @xmath94 as @xmath38 increases , the rate of extinction will remain roughly constant but now the rate of speciation will decrease until a balance is found at @xmath95 . from ( [ e : se_kemax ] ) and ( [ e : se_ksnew ] ) , the corresponding value of @xmath38 is @xmath96 a convenient way to display the crossover in behaviour from small @xmath20 to the noise dominated regime is to consider @xmath0 as a function of . according to ( [ e : se_ninfty ] ) and ( [ e : se_nnew ] ) , this should change from for small @xmath97 to for large @xmath97 . numerical results in support of this prediction
are presented in fig .
[ f : se_crossover ] .
in summary , we have postulated one possible way in which models of macroevolution based on the bak - sneppen approach can be extended to incorporate speciation , extinction and abiotic influences .
the speciation and extinction mechanisms are defined purely in terms of each individual species fitness landscape , irrespective of the total number of species in the system .
nonetheless , the total diversity fluctuates around a constant value @xmath0 , which was termed the _ natural system size _ to stress that it was not arbitrarily chosen .
although the proposed mechanism for speciation , _ ie .
_ a population simultaneously crossing two different fitness barriers , seems appealing , the extinction mechanism is far more heuristic and somewhat unsatisfactory .
a better model might focus on trying to find a more plausible means of extinction , defined in terms of the fitness landscapes .
for instance , the species chosen for evolution might be made extinct if its fitness barrier is below some threshold value .
it may also be possible to place the model on a web structure , and allow the connections themselves to be subject to alteration whenever a species evolves to a new form .
the value of @xmath0 was found to be only weakly dependent upon the average number of connections per species in the system , in agreement with known data .
this leads us to hope that simple models such as ours may be able to reproduce the essential behaviour of real ecosystems .
more realistic models consider the full fitness landscapes rather than just the barriers , but the increased complexity limits the system sizes that can be simulated ( kauffman 1993 ) . a practical step forward might be to reduce known biological principles to simple rules that may be applied to global ecosystems .
we would like to thank prof .
mark newman for useful discussions concerning our model .
* bak * p. and sneppen k. 1993 _ `` punctuated equilibria and criticality in a simple model of evolution '' _ phys .
rev . lett . * 71 * 4083 - 4086 * bak * p. 1997
_ `` how nature works : the science of self - organized criticality '' _ oxford university press * benton * m. j. 1995 _ `` diversification and extinction in the history of life '' _ science * 268 * 52 - 58 * caldarelli * g. , higgs p. g. and mckane a. j. 1998 _ `` modelling coevolution in multispecies communities '' _ preprint adap - org/9801003 * dawkins * r. 1982 _ `` the extended phenotype '' _ oxford university press * flyvberg * h. , sneppen k. and bak p. 1993 _ `` mean - field theory for a simple model of evolution '' _ phys . rev . lett . * 71 * 4087 - 4090 * hall * s. j. and raffaelli d. g. 1993 _ `` food webs - theory and reality '' _ adv
. res . * 24 * 187 - 239 * head * d. a. and rodgers g. j. 1997 _ `` speciation and extinction in a simple model of evolution '' _ phys . rev .
e * 55 * 3312 - 3319 * kauffman * s. a. 1993 _ `` the origins of order '' _ oxford university press * kramer * m. vandewalle n. and ausloos m. 1996 _ `` speciations and extinctions in a self - organising critical model of tree - like evolution '' _ j. phys .
i ( france ) * 6 * 599 - 606 * maynard - smith * j. 1993 _ `` the theory of evolution '' _ cambridge university press * newman * m. e. j. 1997 _ `` a model of mass extinction '' _ j. theor .
biol . * 189 * 235 - 252 * paczuski * m. maslov s. and bak p. 1996
_ `` avalanche dynamics in evolution , growth and depinning models '' _ phys .
e * 53 * 414 - 443 * papoulis * a. 1991 _ `` probability , random variables and stochastic processes '' _ mcgraw - hill * peliti * l. 1997 _ `` an introduction to the statistical theory of darwinian evolution '' _ preprint cond - mat/9712027 * ridley * m. 1993 _ `` evolution '' _ blackwell scientific publications * roberts * b. w. and newman m. e. j. 1996 _ `` a model for evolution and extinction '' _ j. theor . biol .
* 180 * 39 - 54 * sol * r. v. and manrubia s. c. 1996 _ `` extinction and self - organised criticality in a model of large - scale evolution '' _ phys
. rev .
e * 54 * r42 - 45 * sol * r. v. , manrubia s. c. , benton m. and bak p. 1997a _ `` self - similarity of extinction statistics in the fossil record '' _ nature * 388 * 764 - 767 * sol * r. v. and manrubia s. c. 1997b _ `` criticality and unpredictability in macroevolution '' _ phys . rev .
e * 55 * 4500 - 4507 * wilke * c. and martinetz t. 1997 _ `` simple model of evolution with variable system size '' _ phys
e * 56 * 7128 - 7131
in this appendix , the term for external noise that appears in ( [ e : se_stdynoise ] ) is derived .
assuming that @xmath20 is small , noise effects alone will result in @xmath43 taking the mean value of the surrounding square with sides @xmath20 , that is the new term @xmath105 is added to ( [ e : se_evolution ] ) , the expression for @xmath48 for when @xmath85 , to give the total change in @xmath43 at every timestep for @xmath37 . setting this to zero then gives the new steady state equation ( [ e : se_stdynoise ] ) .
.observed values of the natural system size @xmath0 for different @xmath31 , @xmath18 and @xmath20 .
the standard deviation of the fluctuations are given in brackets , which also serve as rough error bars .
the data has been averaged over at least three separate rune of timesteps each .
note that the line for @xmath106 , @xmath107 and @xmath108 appears three times to allow for easier comparison .
[ cols="^,^,^,>,>",options="header " , ] | we describe a simple model of evolution which incorporates the branching and extinction of species lines , and also includes abiotic influences .
a first principles approach is taken in which the probability for speciation and extinction are defined purely in terms of the fitness landscapes of each species .
numerical simulations show that the total diversity fluctuates around a natural system size @xmath0 which only weakly depends upon the number of connections per species .
this is in agreement with known data for real multispecies communities .
the numerical results are confirmed by approximate mean field analysis . 2
the bak - sneppen model was introduced to illustrate the possible role of self - organised criticality in evolving ecosystems ( bak 1993 , 1997 ) .
it is a toy model that describes every species by a single scalar quantity , relating to the expected time before that species evolves to a new form .
interactions in the ecosystem are incorporated by assuming that a species that evolves can alter the time taken for other species to evolve , such as those involved in predator - prey or host - parasite relationships .
the model is said to be _ self - organised critical _ because it approaches a critical state without any apparent need for fine parameter tuning .
consequently , it predicts that extinction events of magnitude @xmath1 should occur with a frequency proportional to @xmath2 , which is consistent with known paleobiological data ( sol 1996 ) .
further evidence has come from analysis of the temporal distribution of extinctions , which exhibits `` 1/f noise '' , also predicted by the model ( sol 1997a ) .
other simple models of macroevolution have now been devised which also claim agreement with the paleobiological data ( peliti 1997 ) .
some of these are believed to be self - organised critical ( sol 1997b ) , although others exhibit power law behaviour via different mechanisms ( roberts 1996 , newman 1997 ) .
all of these models have in common the assumption of a constant system size .
this has been justified by assuming that each species occupies a single _ ecological niche _ , and that if a species is made extinct its niche is immediately filled by a similar , newly emerged species .
the concept of an ecological niche refers to a set of conditions and interactions within the ecosystem that only a single species can satisfy .
however , since the definition of a niche depends upon the other species , the total number of niches should be defined from within the system itself rather than being fixed to some arbitrary constant value for the benefit of computer simulation .
models have been devised which are similar to the bak - sneppen model but allow the total number of species to vary in time .
kramer _
et al . _ introduced a model in which the species are placed onto a branching phylogenetic tree structure , where each species only interacts with its closest relatives ( kramer 1996 ) . depending upon the choice of a parameter , the tree either expands indefinitely or stops growing after a finite time . in the model of wilke _ et al .
_ the system refills after an extinction event at a rate according to a parameter @xmath3 , which is `` the maximal number of species that can be sustained with the available resources '' ( wilke 1997 ) .
however , since the resources are themselves biotic , we believe that any such @xmath4 should be defined from within the system . in this paper , we derive and study a version of the bak - sneppen model in which the probabilities for speciation and extinction are defined purely in terms of the individual species .
nonetheless , the total diversity fluctuates around a natural system size without the need for global control .
in particular , there is no recourse to `` ecological niches '' .
the rules of the model are based on careful considerations of the motion of species on their fitness landscapes .
this is described in sec .
[ s : se_desc ] , and the results of numerical simulations of the model are presented in sec . [
s : se_num ] .
comparisons to real ecosystems are made in sec .
[ s : se_real ] .
approximate analysis of the model is presented in sec .
[ s : se_anal ] which supports and enhances the numerical work . finally , we discuss our results in sec .
[ s : se_disc ] . |
the observational infrastructure enabling the study of solar physics has seen dramatic advances over the past five years . among these , we highlight a few : in 2009 , the two stereo spacecraft ( launched into earth - trailing and earth - leading orbits in 2006 ; @xcite ) passed the quadrature points relative to the sun - earth line .
when combined with earth - perspective viewing , the following years enabled a view of the entire solar surface , for the first time in history showing us the evolution of an entire stellar atmosphere . the atmospheric imaging assembly ( aia ; @xcite ) on board the solar dynamics observatory ( sdo - launched in 2010 ; @xcite ) provides uninterrupted observing of the outer atmosphere of the entire earth - facing side of the sun at a cadence of 12s and a resolution of close to an arcsecond .
combined with magnetography and helioseismology with the helioseismic and magnetic imager ( hmi ; @xcite ) , as well as sun - as - a - star spectroscopy in the euv with the euv variability experiment ( eve ; @xcite ) , this powerful sdo spacecraft sends down well over 1tb / day .
the primary data and higher - level derivatives fill a data archive that now exceeds 7pb and holds over 96% of all data ever taken from space in the domain that focuses on the sun and the physics of the sun - earth connections .
instrumentation with high spatial or spectral resolution is flown on the jaxa - led hinode / solar - b mission ( launched in 2006 ; @xcite ) and the nasa small explorer iris ( interface region imaging spectrograph , launched in 2013 ; @xcite ) that offer images with resolutions between 0.1 and 0.3arcsec , combined with spectroscopy in the visible and ultraviolet .
rhessi @xcite , launched in 2002 , is continuing to provide unique spectroscopic images of the sun at high energy .
these space - based instruments provide important access to the domain from photosphere to corona , critically complemented by ground - based telescopes and their instrumentation , as well as by rocket experiments such as hi - c @xcite . in the optical domain , the 1.6 m new solar telescope ( nst , @xcite ) at big bear in the u.s.a . and the 1.5-meter german gregor solar telescope at tenerife in spain @xcite have been in regular operation since 2012 .
the 1-m new vacuum solar telescope ( nvst , @xcite ) at fuxian lake in southwest china is also in regular operation recently .
crisp ( crisp imaging spectro - polarimeter ) at the swedish 1-m solar telescope ( sst ; ) reached 0.13arcsec spatial resolution and high polarimetric sensitivity aided by post - processing .
all these telescopes have capacities close to their diffraction limit due to advanced designs and excellent seeing conditions .
we have seen glimpses into very high - resolution ground - based coronal observing as well with , for the first time , synoptic observations of coronal stokes polarimetry in the near infrared by the coronal multichannel polarimeter ( comp , @xcite ) telescope . in the radio domain ,
the recently commissioned chinese spectral radioheliograph ( csrh ; ) at mingantu in inner mongolia of china ( renamed as mingantu ultrawide spectral radioheliograph - muser ) is a radioheliograph operating with the widest frequency range ever reached from 400mhz to 15ghz , with a high temporal , spatial , and spectral resolution .
recent non - solar dedicated radio arrays such as the murchison wide - field array ( mwa , @xcite ) and the low - frequency array for radio astronomy ( lofar , @xcite ) have obtained spectroscopic solar imaging at metric and lower frequencies .
the recently upgraded karl g. jansky very large array ( evla ) has provided solar radio dynamic imaging spectroscopy of type iii bursts at decimeter wavelengths @xcite .
even the millimeter domain and beyond is seeing major advances with the atacama millimeter / sub - millimeter array ( alma ; @xcite ) observatory coming on line for solar observing as it continues to complete its construction phase .
the observations of tens of thousands of sun - like stars by the nasa kepler satellite is offering yet more insights into the physics of the sun , ranging from an improved understanding of internal structure and dynamics ( with asteroseismic techniques ) to a view of rare , extremely energetic flares .
kepler data , with ground - based follow - up studies , suggest that solar flares may occur with energies up to several hundred times higher than observed directly in the past half century with space - based instrumentation @xcite .
the volume of information that needs to be processed by solar researchers is increasing rapidly . in terms of data to be analyzed , we have definitely reached the petabyte era .
this is true for observational data in the archives of sdo , but also in computer experiments in which single snapshot data dumps of the advanced codes can exceed a tb .
a tendency towards open data policies means that we have ever more access to large volumes and a daunting diversity of data .
that complicates finding , processing , and integrating data . infrastructure support for , for example ,
the virtual solar observatory ( vso ) , solarsoft idl ( sswidl ) , and the heliophysics events knowledgebase and registry ( hek , her ) are critically important to enable efficient utilization of the growing data diversity and volume .
the community lags in strategic thinking about these meta - infrastructural elements , both where current support and future expansion or replacement are concerned .
there has also been a recent movement towards open - source data analysis software , with the development of the sunpy solar analysis environment in python @xcite .
a similar flood of information is found with scientific publications , which exceed 2,000 refereed publications per year ( see above ) . here
, the support infrastructure of ads is of critical value .
the wide diversity of journals in which the works of colleagues are published requires subscription access to many publications , at costs that are increasingly hard to bear for relatively small research groups ; here , preprint services such as arxiv and maxmillennium play significant roles in making the community aware of what is going on .
`` living reviews '' as offered by the free on - line journal living reviews in solar physics enable new researchers to understand the context of their work and established researchers in one sub - specialty a quick introduction into adjacent areas by their peers . among the difficulties the solar activity community also faces
is that many solar and inner - heliospheric events are studied by different groups and published in different journals .
finding studies on a particular solar region of interest is hampered by inconsistent use of the characterizing spatio - temporal coordinates of events which may be found in abstract , main text , tables , appendices , and sometimes only marked in figures that are not machine - readable .
the iau adopted a standard convention for this in 2009 ( @xcite ) in its sol standard ( short for solar object locator ) .
its use is encouraged by journal editors including those of solar physics , the astrophysical journal , and the journal of geophysical research .
broad usage of the sol standard would enable computer searches of related publications , enabling researchers to put new ( meta-)studies in broader contexts .
for the highlights touched upon in this report , we opted for two criteria to identify topics of interest .
one is to mention specific areas of note in the opinion of the organizing committee that may be new developments , are highly specialized yet significant , may concern new instrumentation or methods , or are otherwise deemed to be developments that may grow to see more activity in terms of publications in the near future . the other criterion we applied is to be guided towards topics of frequent activity by the community itself by looking at the most cited works . such a selection does introduce a bias toward the papers published early in the period reviewed , of course , but our purpose is not to identify the most - cited works per se , but rather to find the dominant themes within the set of these works that apparently resonate strongly within the community already within the 3-y window from which they are selected .
we searched for the most - cited refereed publications from ads with the terms `` solar.activity '' , `` coronal.mass.ejection '' , `` solar flare '' , `` solar prominence '' , or synonym(s ) in their abstracts in the period 2012 - 2014 . within this set , we identify the following themes ( sorted alphabetically ) : active - region magnetic fields ; coronal thermal structure ; coronal seismology ; solar flares and eruptions ; and the sun - in - time related aspects of long - term solar variability including cosmic - ray modulation .
we close with a collection of developments , discoveries , and surprises .
the photospheric magnetic field of active regions forms the foundation of the overlying atmosphere .
its evolution through emergence , displacement , and submergence of flux is key to driving eruptive and explosive events in the corona and into the heliosphere . until recently ,
generally only line - of - sight magnetic field maps were available for this work .
nowadays , regular vector - magnetic determinations from the observed polarization signals are available from sources that include hinode , solis , and sdo / hmi .
the sensitivity of such vector field maps allows the detection of lasting changes in the photospheric field when comparing pre- to post - flare observations ( _ e.g. _ * ? ? ?
temporal resolution is so good that coronal events can be tightly bracketed to try to understand the causes of flares and eruptions , as well as the changes in energy and helicity involved ; for example , @xcite analyze a series of nonlinear force - free field models for the evolution in the energy of an active region around the time of a major eruption .
the increasing availability of vector - magnetic data enables statistical studies on the properties of active regions and their activity heretofore possible only on line - of - sight magnetograms .
for example , @xcite use a data base of vector field maps of over 2,000 active regions to train a machine - learning algorithm to attempt forecasting of large flares .
@xcite review 3,000 vector field maps of some 60 regions to compare estimates of free energy with flare rates .
@xcite analyze a sample of nearly 200 coronal mass ejections to reveal that flux , twist , and proxies for free energy tend to set upper limits to the speed of cmes emanating from the active regions studied .
@xcite review hundreds of vector magnetograms of a sample of 80 active regions to study helicity and twist parameters to test for the influence of coriolis forcing .
one long sought - after goal of flare and cme physics is to use models of the solar atmospheric field to understand why field configurations destabilize and under what conditions destabilization begins and proceeds . in a `` meta - analysis '' review , @xcite discuss recent developments , including the use of 3d field extrapolations that suggest that topological structures ( notably null points and hyperbolic flux tubes ) may be involved in the triggering and generally as tracers of reconnection processes early on .
but even as the availability of vector - magnetic field maps becomes routine , the realization is growing that by themselves they appear insufficient to provide generally adequate lower - boundary conditions to understand the solar atmospheric activity .
for example , a group of modelers using a variety of non - linear force - free field algorithms concludes in a series of studies ( see , _ e.g. _ * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) that snapshot vector - magnetic maps do not suffice to obtain a reliable coronal field model with accurate energy or helicity measurements , with effects of instrumental resolution and field of view , as well as the model geometry ( cartesian vs. spherical ) all compounding the problems .
new developments include the use of coronal loop observations to guide non - potential field models either from a single perspective as is possible currently ( _ e.g. _ * ? ? ? * ) or by using stereoscopic data from existing or future space- based instrumentation such as stereo , solar orbiter , and sdo ( _ e.g. _ * ? ? ?
* ; * ? ? ?
methods to follow the evolution of active - region fields based on data driving are also being developed , using the uninterrupted stream of ( vector ) magnetograms now available from space - based platforms ( _ e.g. _ * ? ? ?
* ; * ? ? ?
* based on the mhd - like magnetofrictional approximation ) , even reaching up to global mhd field models from near the solar surface into the heliosphere @xcite .
others are developing mhd methods to study cme initiation based on observed surface field evolution ( _ e.g. _ * ? ? ? * ) .
fundamental difficulties with these emerging methods include the difficulty in measuring the transverse field in areas of relatively low flux densities , the intrinsic 180-degree ambiguity in the field direction given a magnitude for the transverse component , the need to constrain the electric field to drive the model s evolution , and ultimately the quantitative comparison with solar observables to determine the verisimilitude of the model fields .
even as our ability to observe and process rapidly growing data volumes on active region fields increases , we remain puzzled by the sun s atmospheric magnetic field : we have yet to understand how large amounts of energy can sometimes be stored to eventually be explosively converted into flares and cmes , while in other cases the energy is either not stored or is not explosively released . for a recent review of our understanding of the magnetic field in the solar atmosphere , and the variety of methods used to observe and study it , we refer to , and references therein .
the x - ray solar corona is made of complex arrays of magnetic flux tubes anchored on both sides to the photosphere , confining a relatively dense and hot plasma .
this optically thin plasma is almost fully ionized , with temperatures above 1mk , and emitting mostly in the extreme uv to x - rays with intensities proportional to the square of its density .
significant progress was made in studying the physical and morphological features of coronal loops in a series of very successful euv and x - ray solar missions , since the first observational evidence of the presence of coronal loops provided by rocket missions in the mid-1960s @xcite . in recent years
, several solar missions were launched and started producing spectacular images and data in different spectral ranges .
the high spatial and temporal resolution of these instruments and the complementarity between these data sets , poses new challenges to understand the heating and dynamics of coronal loops .
the launch in 2010 of the solar dynamics observatory ( sdo , * ? ? ?
* ) allows continuous observation of the whole sun with high temporal and spatial resolution with aia , eve , and hmi .
in particular , aia images span at least 1.3 solar diameters in multiple wavelengths , at about 1arcsec in spatial resolution and at a cadence of about 10seconds .
coronal loops observed by aia have been found to be highly variable and highly structured in space , time , and temperature , challenging the traditional view of these loops as isothermal structures and favoring the case for multi - thermal cross - field temperature distributions . because the thermal conductivity is severely reduced in the directions perpendicular to the magnetic field
, a spatially intermittent heating mechanism might give rise to a multi - threaded structure in the internal structure of loops .
one possible example of such an intermittent heating process is mhd turbulence , which is expected to produce fine scale structuring within loops all the way beyond the resolution of current observations @xcite .
a recent study by @xcite combining spectroscopic data from the euv imaging spectrometer ( eis , * ? ? ? * ) aboard hinode and sdo / aia images , shows that most of their loops must be composed of a number of spatially unresolved threads .
more recently , the sounding rocket mission high - resolution coronal imager ( hi - c , * ? ? ?
* ) , achieved an unprecedented spatial resolution 0.2arcsec in euv images .
@xcite find that the finely structured corona , down to the 0.2 resolution , is concentrated in the moss and in areas of sheared field , where the heating is intense .
this result suggests that heating is on smaller spatial scales than aia and that it could be sporadic .
these results are consistent with differential emission measure ( dem ) analysis that study the distribution of temperature across loops .
@xcite present a systematic study of the differential emission measure distribution in active region cores , using data from eis and the x - ray telescope ( xrt ) aboard hinode .
their results suggest that while the hot active region emission might be close to equilibrium , warm active regions may be dominated by evolving million degree loops in the core .
more recently , @xcite used xrt and eis data as well as images from sdo / aia , and found that cooler loops tend to have comparatively narrower dem widths . while the dem distribution of warm loops could be explained through bundles of threads with different temperatures ,
cooler loops are consistent with narrow dems and perhaps even isothermal plasma .
the authors then speculate that warm , multi - thermal , multi - threaded loops might correspond to plasma being heated , while cool loops are composed of threads which have had time to cool to temperatures of about a million degrees , thus resembling a single isothermal loop .
the interface region imaging spectrograph ( iris , * ? ? ?
* ) was launched in 2013 , and provides crucial information to understand coronal heating , by tracing the flow of energy and plasma from the chromosphere and transition region up to the corona .
iris obtains high resolution uv spectra and images with high spatial ( 0.33arcsec ) and temporal ( 1s ) resolution .
recent iris observations show rather fast variations ( 20 - 60s ) of intensity and velocity on spatial scales smaller than 500 km at the foot points of hot coronal loops @xcite .
these observations were interpreted as the result of heating by electron beams generated in small and impulsive heating events ( the so - called coronal nanoflares ) .
theoretical models of coronal heating have been traditionally classified into ac or dc , depending on the time scales involved in the driving at the loop foot points : ( a ) ac or wave models , for which the energy is provided by waves at the sun s photosphere , with timescales much faster than the time it takes an alfvn wave to cross the loop ; ( b ) dc or stress models , which assume that energy dissipation takes place by magnetic stresses driven by slow foot point motions ( compared to the alfvn wave crossing time ) at the sun s photosphere .
although these scenarios seem mutually exclusive , two common factors prevail : ( i ) the ultimate energy source is the kinetic energy of the sub - photospheric velocity field , ( ii ) the existence of fine scale structure is essential to speed up the dissipation mechanisms invoked @xcite . for a coronal heating mechanism to be considered viable
, the input energy must be compatible with observed energy losses in active regions , estimated by @xcite to be @xmath0 .
@xcite used high - resolution observations of plage magnetic fields made with the solar optical telescope aboard hinode to estimate the vertical poynting flux being injected into the corona , obtaining values of about @xmath1 , which suffices to heat the plasma .
coronal seismology , like terrestrial seismology and traditional helioseismology , provides a means of probing the background state of the medium through which the waves propagate .
although the corona is not hidden from view like the interiors of earth and sun , it is very difficult to directly measure plasma and magnetic properties such as density @xmath2 , magnetic field @xmath3 , or transport coefficients .
observations of oscillations in the coronal plasma can potentially provide powerful constraints on these quantities , for example through the alfvn velocity @xmath4 ( _ e.g. _ , see the _ living review _ of * ? ? ?
but unlike the solar interior and terrestrial examples where waves are but perturbations on the background state , the coronal seismic waves are crucially important to the energy balance of their host medium .
coronal heating and solar wind acceleration are widely thought to result at least in part from waves @xcite .
an overview of coronal seismology as of 2012 is provided by @xcite .
the last 58 years have seen an explosion in coronal wave studies due to the advent of new instrumentation , such as the ground - based coronal multichannel polarimeter ( comp , * ? ? ? * ) and the space - based atmospheric imaging assembly ( aia ) on the solar dynamics observatory ( sdo ) . both have revealed ubiquitous alfvn - like ( i.e. , transverse to the magnetic field ) coronal oscillations ( * ? ? ?
* respectively ) , though the interpretation of their exact natures alfvn or kink is controversial @xcite .
the term `` alfvnic '' is commonly used to encompass both wave types . in either case , the increased resolution of the aia observations @xcite has allowed the detection of sufficient oscillatory power to potentially power both the corona and solar wind @xcite , though precise mechanisms are not currently known with certainty .
the concentration of oscillatory power in the few - minute period band , and in particular at around 5-minutes , strongly suggests a link with the sun s internal @xmath5-mode global oscillations .
consistency in the estimation of alfvn speed between seismic techniques and magnetic field extrapolation is confirmed by @xcite using aia data from a flare - induced coronal loop oscillation .
time - distance techniques applied to comp observations indicate a preponderance of outward propagating waves over inward propagation , even in closed loop structures , suggesting _ in situ _ dissipation ( or mode conversion ) on a timescale comparable to the alfvn crossing time @xcite . disentangling alfvn and kink waves
is addressed in depth by @xcite . in the magnetically structured solar atmosphere , the only true alfvn wave is torsional , and there is considerable interest in identifying these in observations because of the amount of energy they could potentially contribute to the outer atmosphere . however , being incompressive , alfvn waves are not seen in intensity , and torsional alfvn waves are also difficult to detect in doppler @xcite .
recently though , 0.33-arcsec high - resolution observations of the chromosphere and transition region ( tr ) with nasa s interface region imaging spectrograph ( iris ) , coordinated with the swedish solar telescope , have revealed widespread twisting motions across quiet sun , coronal holes , and active regions alike that often seem to be associated with heating @xcite .
this must presumably extend to the corona as well .
dissipation of alfvn waves in the solar atmosphere has long been thought to rely largely on the generation of alfvn turbulence through nonlinear interaction between counter - propagating waves @xcite .
observations with comp showing enhanced high - frequency power near the apex of coronal loops @xcite possibly supports this view .
the presumed acceleration of the solar wind by alfvn turbulent energy deposition poses the challenge of identifying and explaining counter - propagating alfvn waves in open magnetic field regions required to produce that turbulence . @xcite
confirm the presence of these counter - propagating waves using comp .
these observations also provide evidence of a link to the @xmath5-mode spectrum , which presumably relies on magnetoacoustic - to - alfvn mode conversion occurring in the lower atmosphere @xcite .
the solar chromosphere is where most of the energy of a solar flare is dissipated and radiated in bright linear structures called flare ribbons , and recent work on this topic has been dominated by observations from iris .
the high spectral resolution available with iris shows complex spectral line profiles in the si iv transition region line at 1394 and 1403 sometimes with two or three time - varying gaussian components within one iris spatial pixel @xcite .
the hot `` coronal '' line of fe xxi ( 1354 ) on the other hand shows only a single strongly blue - shifted component @xcite originating in compact ribbon sources at the beginning of a flare , with no hot `` stationary component '' present .
the presence of this line also demonstrates the high temperatures reached by the chromosphere in flares as deduced from hinode / eis observations @xcite . with eis
, we see 1.5 - 3mk redshifts @xcite confirming earlier reports , as well as significant non - thermal broadening .
this means that the momentum - conserving condensation front that is produced by flare heating and paired with the evaporation flow contains hot plasma . in turn
this implies that the condensation front originates relatively high in the chromosphere , otherwise such high temperatures would not be possible in the standard ( electron - beam - driven ) model of flares .
this is somewhat at odds with recent measurements of element abundance @xcite which look more photospheric than coronal , suggesting that up - flowing evaporated material comes from low down in the chromosphere , below where the normal fractionation by first ionization potential sets in .
flare optical ( or white - light
wl ) emission continues to be difficult to observe and difficult to explain .
optical foot points characterized using 3-filter observations with hinode / sot @xcite could be explained by modest temperature increases of the photospheric black body .
the other main proposed radiation mechanism is recombination emission , and flare continuum in the near uv ( beyond the balmer edge ) observed with iris has an intensity consistent with this @xcite , but the tell - tale balmer jump has not been seen .
co - spatial hard x - ray ( hxr ) and white - light flare sources have been observed using rhessi and sdo / hmi , and require that both emissions are produced a few hundred kilometres above the photosphere @xcite , at a height corresponding to the temperature minimum region , and beyond the range expected for hxr - emitting electrons arriving from the corona ( unless the chromosphere is under - dense compared to expectations ) .
it may be possible to generate optical emission from the temperature minimum region , for example by modest heating due to ion - neutral damping @xcite , but the presence of non - thermal electrons in this plasma is harder to explain . also indicating the flare s impact on the dense lower atmosphere , there have been many more reports of `` sun quakes '' flare seismic emission ( e.g. * ? ? ?
* ; * ? ? ?
* ) but the mechanical driver remains uncertain ; there are correspondences with either hxr sources ( pressure pulses from electron - beam driven shocks ) or magnetic transients ( lorentz forces ) in some but not all cases .
high - resolution ground - based flare observations using the new solar telescope at bbso show extraordinary fine structure , on a sub - arcsecond scale , in flare ribbons and footprints @xcite , setting the scene for future observations with dkist .
examination of the changes in photospheric vector field occurring at the flare impulsive phase by @xcite using the hmi on sdo suggests strong , permanent and abrupt variations in the vertical component of the lorentz force at the photosphere consistent with a downward ` collapse ' of magnetic loops , and changes in the horizontal component mostly parallel to the neutral line in opposite directions on each side , indicating a decrease of shear near the neutral line .
the brighter the flare ( as expressed in the goes class ) , the larger are both the total ( area - integrated ) change in the magnetic field and the change of lorentz force @xcite , although @xcite found that indicators of magnetic non - potentiality ( e.g. rotation , shear and helicity changes ) are more closely associated with flares see for example the study of flare - associated rotating sunspots by @xcite . in the corona , imaging and spectroscopic observations show compelling evidence for plasma flows suggestive of those expected around a coronal reconnection region ( e.g. * ? ? ?
* ; * ? ? ?
* ) . the growing number of observed flares in archives and the increased coverage of flares in wavelength space and in domains from surface to heliosphere is enabling an improved assessment of energy budgets .
for example , @xcite quantify energies of an ensemble of large , eruptive flares to find , among others that it appears that the energy in accelerated particles during the initial phases of the flare suffice to supply the energy eventually radiated in the flare across the spectrum , and that that total energy is statistically just under the bulk kinetic energy in associated coronal mass ejections .
the central problem in solar flare theory remains the acceleration of the non - thermal electrons required to explain observed chromospheric hxr sources .
observations with rhessi and sdo show that coronal electron acceleration can be very efficient ; @xcite find that essentially all electrons in a coronal source of density a few times @xmath6@xmath7 are energized to above around 10kev .
kappa distributions , which are found to be a better fit than a standard thermal plus non - thermal distribution in coronal hxr sources @xcite , are shown to arise naturally in an acceleration region when there is a balance between diffusive acceleration and collisions , in the absence of significant escape from the acceleration region @xcite . in the electron - beam model of a flare
, electrons must of course escape the corona to produce the chromospheric hxr sources , and the number flux requirements have always been somewhat problematic .
this may be alleviated if electrons are boosted by wave - particle interactions in the corona ; a quasi - linear simulation of coronal electron propagation shows that wave - particle interaction with the high phase - velocity langmuir waves generated by density inhomogeneities can accelerate beam electrons to higher energies , reducing the requirement on electron flux at energies of a few tens of kev by up to a factor ten .
study a model in which electrons are re - accelerated in the chromosphere , concluding that this also reduces demands on putative electron beam fluxes ( however requirements on the flare chromosphere energy source , not directly addressed in this model , remain the same ) .
chromospheric ( re- ) acceleration models may also produce electron angular distributions which are more isotropic , consistent with the angular distributions inferred from inversion of rhessi mean electron flux spectra , accounting for photospheric x - ray albedo , by @xcite .
a completely different view by @xcite is that the electron energization in flares takes place in a parallel electric field that develops close to or in the chromosphere , in a region of anomalous resistivity , if energy is transported alfvnically specifically by inertial alfvn waves .
however , the orthodoxy remains that energy transport is by electron beams , and this is now being tested against observations using beam - driven radiation hydrodynamics codes , the output of which can be compared with , for example , iris @xcite and eve and aia data , so far with mixed success .
the understanding of the initiation and evolution of coronal mass ejections ( cmes ) has tremendously profited from the combination of coronagraphic observations with high cadence imaging in the euv together with the multi - perspective view provided by the stereo mission , as well as from increasingly sophisticated mhd and thermodynamic modeling ( for reviews see * ? ? ?
* ; * ? ? ?
magnetic flux ropes play a key role in the physics of cmes .
but there is a long debate whether flux ropes are pre - existing or formed during the eruption .
@xcite observed the formation of a flux rope during a confined flare in high - cadence sdo / aia euv imagery . within hours after its formation , the flux rope became unstable and erupted resulting in a cme / flare event .
for other cmes , for example those associated with quiescent prominence - cavity systems , a variety of observations indicate a pre - existing flux rope which may erupt bodily as a cme ( see e.g. figure 12 of @xcite ) .
@xcite synthesized 16 years of coronagraphic and euv observations with mhd simulations , and found that flux ropes are a common structure in cmes ; in at least 40% a clear flux rope structure could be identified .
in addition , they established a new `` two - front '' morphology consisting of a faint front followed by diffuse emission and the bright cme leading edge .
the faint front is suggestive of a wave or shock front driven by the cme .
the high - cadence six - passband sdo / aia euv imagery allows to perform differential emission measure ( dem ) analysis on solar flares and cmes to study their multi - thermal dynamics .
it was shown that the cme core region , typically identified as the embedded flux rope , is hot ( 8 10 mk ) indicative of magnetic reconnection being involved .
in contrast , the cme leading front has temperatures similar to the pre - eruptive corona but of higher densities suggesting that the front is a result of compression of the ambient coronal plasma .
the hot flux rope is a key indicator to the physical processes involved in the early acceleration phase of the cme @xcite .
the environment of cmes is important for the development of non - radial propagation .
@xcite report that during solar minimum conditions cmes originating from high latitudes can be easily deflected toward the heliospheric current sheet , thus eventually becoming geo - effective .
@xcite showed that coronal holes nearby the cme initiation site can cause strong deflections of cmes .
modeling of the initiation of cmes continues to provide insights into the various forces and mechanisms that may be involved : initiation may involve the kink instability @xcite , sunspot rotation , reduction of tension of the overlying field @xcite , torus instability @xcite , and the breakout process @xcite . which dominates under which conditions and how commonly these occur remain topics of future work . since their discovery by the soho / eit instrument about 15 years ago
, there has been a vivid debate about the physical nature of large - scale euv waves , i.e. whether they are true wave phenomena or propagating disturbances related to the magnetic restructuring due to the erupting cme . in the recent years
, there has been tremendous progress in the understanding of these intriguing phenomena thanks to the unprecedented observations available , in particular the high - cadence euv imagery in six wavelengths bands by sdo / aia combined with the stereo multi - point view which allowed for the first time to follow euv waves in full - sun maps @xcite .
there seems now relatively broad consensus that large - scale euv waves are often fast - mode magnetosonic waves ( of large amplitude or shocks ) , driven by the strong lateral expansion of the cme ( see reviews by * ? ? ?
* ; * ? ? ?
a number of detailed case studies revealed that the cme lateral front and the euv wave appear originally co - spatial .
but when the lateral cme expansion slows down , the euv wave decouples from the driver and then propagates freely , adjusting to the local fast - mode speed of the medium ( e.g. * ? ? ?
* ; * ? ? ?
three - dimensional thermodynamic mhd modeling of well observed euv waves also supports these findings , showing the outer fast - mode euv wave front followed by another bright front indicating the cme component @xcite .
statistical studies of euv waves based on sdo / aia @xcite and stereo / euvi data @xcite revealed euv wave speeds that range from close to the fast magnetosonic speed in the quiet corona to values well above , the fastest ones exceeding 1000 km s@xmath8 .
@xcite showed that at least half of the euv waves under study show significant deceleration , and a distinct anti - correlation between the starting speed and the deceleration , providing further evidence for a freely propagating fast - mode wave .
the association rate of euv waves with type ii bursts , which are indicative of shock waves in the solar corona , may be as high as 50% @xcite .
detailed case studies provided a number of further characteristics suggestive of the wave nature , such as reflection and refraction of euv waves at coronal holes and active regions , transmission into coronal holes as well as the initiation of secondary waves by the arrival of the wave at structures of high alfvn speed @xcite and for one case , @xcite have evaluated the euv wave s initial energy using a blast - wave approximation , to be around 10% of that of the associated cme .
@xcite discovered quasi - periodic fast - mode wave trains within a large - scale euv wave with a periodicity of 2 min , running ahead of the laterally expanding cme flanks .
@xcite presented the first simultaneous observations of the propagation of a large - scale euv wave and an h@xmath9 moreton wave , showing that the wave fronts evolve co - spatially indicating that they are both signatures of a fast magnetosonic wave pulse .
the stereo mission , often in combination with soho / lasco , offers observations of cmes all the way from their origin on the sun , and of their propagation in interplanetary space to beyond 1au from outside the sun - earth line .
these data combined with a multitude of other in - situ space missions have been vividly used to connect remote sensing cme observations to the field and plasma data observed by in - situ spacecraft , to constrain models of interplanetary cme propagation , to study cme - cme interaction , and to forecast cme arrival times and speeds with the ultimate aim improving the prediction of their geo - effectiveness .
@xcite and @xcite tracked a flux rope all the way from its solar origin to its in - situ signatures at 1au using the stereo secchi euv imagers , coronagraphs and wide - angle heliospheric imagers .
they establish that the cavity in the classic three - part cme is the feature that becomes the magnetic cloud , implying material ahead of the cavity is piled - up material from the corona or the solar wind .
modeling of the interplanetary propagation of cmes makes use of empirical , analytic and numerical approaches .
the analytical drag - based model " ( dbm ) is based on the hypothesis that the lorentz forces driving a cme eruption ceases in the upper corona and that beyond a certain distance the interplanetary cme ( icme ) dynamics is governed solely by the interaction of the icme and the ambient solar wind plasma @xcite . from the observational side , a variety of reconstruction methods have been developed and applied to the heliospheric imager data including one- as well as two - spacecraft ( stereoscopic ) observations and inclusion of in - situ data and radio type ii bursts to better constrain the propagation direction , distance and speed profile of cmes in interplanetary space @xcite .
these efforts result in comparable typical uncertainties in the cme arrival time of about half a day @xcite .
studies using stereo heliospheric imager data and in - situ plasma and field measurements established that the interaction of cmes in the inner heliosphere , due to a faster cme launched after a slower one , seems to be a common and important phenomenon @xcite . the interaction process may cause deflection or merging of cmes , and either deceleration or acceleration of merged cme fronts ( including heating and compression ) .
@xcite , reporting in a fast cme causing an extreme storm , speculate that the interaction between two successively launched cmes resulted in the extreme enhancement of the magnetic field of the ejecta that was observed in - situ near 1au . the unusually deep and temporally - extended activity minimum between sunspot cycles 23 and 24 , followed by a slowly rising and low amplitude cycle 24 , has led to renewed interest in the underlying causes of solar cycle fluctuations , including grand minima .
much attention has focused on the so - called babcock - leighton solar cycle models , in which the regeneration of the solar surface dipole takes place via the decay of active regions .
most extant versions of these dynamo models are geometrically ( axisymmetric ) and dynamically ( kinematic ) simplified , yet they do remarkably well at reproducing many observed solar cycle characteristics ( see , _ e.g. _ , karak _ et al .
_ 2014 , and references therein ) .
explanations for the extended cycle 23 - 24 minimum and low amplitude cycle 24 have been sought in terms of variations in the meridional flow expected to thread the solar convection zone ( upton and hathaway 2014 ) , and patterns of active region emergence and associated feedback on surface flows ( cameron _ et al .
_ 2014 ; jiang _ et al .
these successes of the babcock - leighton modelling framework have however been challenged by helioseismic measurements ( zhao _ et al .
_ 2013 ; schad _ et al .
_ 2013 ) indicating that the meridional flow within the convection zone has a far more complex cellular structure than assumed in the majority of these mean - field - like solar cycle models .
possible avenues out of this conundrum are being explored ( see , _
e.g. _ , hazra _ et al .
_ 2014 ; belucz _ et al .
_ 2015 ) .
much effort has also been invested in implementing various form of data assimilation schemes in dynamo models , with the aim of achieving improved forecasting of the amplitude and timing of future sunspot cycles . at this point in time
no existing dynamo model - based forecasting scheme has done significantly better than the known precursor skill of the solar surface magnetic dipole moment at times of cycle minima , nonetheless progress is likely forthcoming in this area .
global magnetohydrodynamical simulations of solar convection have also progressed rapidly in recent years , with many research groups worldwide now running simulations producing large - scale magnetic fields undergoing polarity reversals ( _ e.g. _ masada _ et al .
_ 2013 ; nelson _ et al .
_ 2013 ; fan and fang 2014 ; passos and charbonneau 2014 ; warnecke _ et al .
_ 2014 ) . due to computing limitations all these simulations
run in parameter regimes still far removed from solar interior conditions .
nonetheless , many are producing tantalizingly solar - like features , including rotational torsional oscillations ( beaudoin _ et al .
_ 2013 ) equatorward propagation of activity `` belts '' ( kpyl _ et al .
_ 2012 ; warnecke _ et al .
_ 2014 ; augustson _ et al .
_ 2015 ) cyclic in - phase magnetic modulation of convective energy transport ( cossette _ et al .
_ 2013 ) and grand minima - like interruptions of cyclic behavior ( augustson _ et al .
one particularly interesting feature is the spontaneous production of magnetic flux tube - like structures within the convection zone , as reported in nelson _
these were found to rise to the top of the simulation domain , partly through magnetic buoyancy , while maintaining their orientation in a manner compatible with hale polarity laws ( nelson _ et al .
this has revived the idea that dynamo action could be wholly contained within the solar convective envelope , rather than relying on the tachocline for the formation and storage of the magnetic flux ropes eventually giving rise to sunspots .
major efforts have also taken place in reinterpreting and reanalyzing historical observations of magnetic activity .
noteworthy in this respect are the analyses of tilt angle patterns for bipolar magnetic regions ( see pavai _
_ 2015 , and references therein ) , and reanalysis of polar faculae data by munoz - jaramillo _ et al . _
( 2012 ) . of particular importance
is the recently completed revision of the international sunspot number ( ssn ) time series .
ssn values for the period 1947-present now account for a discontinuity in the manner of counting spot groups having occurred at the locarno reference station ( clette _ et al .
correcting for this leads to a significant ( @xmath10% ) decrease in ssn values during the space era .
consequently , reconstructions of solar activity into the distant past using the ssn as a backbone to extrapolate space - borne measurements will need to be reassessed .
radionuclides generated by the atmospheric impact of galactic cosmic rays ( gcrs ) provide a crucial proxy for the evolution in the sun s activity @xcite on time scales longer than a few years or a decade , depending on the radionuclide and its deposition in terrestrial natural archives .
new ice core data on @xmath11be and tree ring data on @xmath12c have been combined to provide better understanding of climate impacts on these records : a joint analysis of composite tree ring data with ice cores from greenland and antarctica have enabled the separation of the common signal ( assumed to be dominated by solar and heliospheric variability ) from terrestrial variability @xcite . from this , we now have 94 centuries of data on a proxy for solar activity . but translating that proxy into details of solar activity that may drive space weather and terrestrial climate remains a challenge , as reviewed by , _
e.g. _ . and
then , of course , there were numerous surprising realizations and discoveries , in the real world as much as in the rapidly growing virtual world ) .
we mention merely a small sampling in no particular order : a weak solar cycle following an uncommonly low and long solar minimum @xcite ; a series of x - class flares from ar12192 none of which were associated with a cme , contrasting with statistics to date @xcite ; use of a sun - grazing comet to probe the high corona and its connection to the innermost heliosphere @xcite ; an extremely large amount of dense , cool plasma falling back onto the sun following a massive filament eruption providing a close - up example of distant accretion processes @xcite ; reports of enormously energetic flares from what would appear to be sun - like stars @xcite and potential evidence for strong sep events associated with very energetic solar flaring from @xmath12c records albeit without obvious auroral counterparts ; the successful creation of realistic looking sunspots in the computer @xcite ; the revision of sunspot numbers that suggests no long - term increase in solar activity occurred over the past few hundred years @xcite ; radiative magneto - convective simulations have reached resolution scales of a few kilometers , and suggest comparable energy densities in magnetic and kinetic reservoirs @xcite ; iris observations uncovering rapidly - evolving low - lying loops at transition region temperatures , heretofore inferred from emission measure studies but never yet observed @xcite ; a new model was proposed for coronal heating based on magnetic gradient pumping @xcite ; non - potential field models for a continuously - driven corona over a 16-year period was achieved @xcite ; a solar eruption in july of 2012 that would likely have powered a century - level extreme geomagnetic storm , as for the carrington - hodgson flare of 1859 , had it enveloped earth @xcite ; the realization that stokes theorem combined with the induction equation could explain why polar fields should be a good indicator for the strength of the next sunspot cycle @xcite ; the simulation of a sequence of homologous cmes and demonstration of so - called `` canniballistic '' behavior @xcite ; the discovery of nested toroidal line - of - sight flows and `` lagomorphic '' coronal polarimetric signatures within coronal cavities indicating the presence of magnetic flux ropes @xcite ; rapidly rotating magnetic structures ( `` magnetic tornadoes '' ) have been identified , which provide a channel of energy and twist from the solar surface to the corona @xcite ; and the x - class flare sol2014 - 03 - 29 t made history by becoming `` the best - observed flare of all time '' ( http://www.nasa.gov/content/goddard/nasa-telescopes-coordinate-best-ever-flare-observations/#.vhj4trq9yow[according to nasa ] ) as the ground - based dunn solar telescope and the space - based iris , rhessi , and sdo all observed it in detail .
201 # 1isbn # 1#1#1#1#1#1#1#1_#1_#1*#1*#1#1#1#1#1#1#1#1#1#1#2#1#1#1#1#1_#1_#1#1#1*#1*#1#1#1_#1_#1#1#1#1#1#1#1#1#1#1#1 # 1http://dx.doi.org/#1[]:#1 # 1http://arxiv.org / abs/#1[]:#1 # 1http://adsabs.harvard.edu / abs/#1[]:#1#1#1#1#1#1#1#1#1#1#1#1#1 | after more than half a century of community support related to the science of `` solar activity '' , iau s commission 10 was formally discontinued in 2015 , to be succeeded by c.e2 with the same area of responsibility . on this occasion
, we look back at the growth of the scientific disciplines involved around the world over almost a full century . solar activity and fields of research looking into the related physics of the heliosphere
continue to be vibrant and growing , with currently over 2,000 refereed publications appearing per year from over 4,000 unique authors , publishing in dozens of distinct journals and meeting in dozens of workshops and conferences each year .
the size of the rapidly growing community and of the observational and computational data volumes , along with the multitude of connections into other branches of astrophysics , pose significant challenges ; aspects of these challenges are beginning to be addressed through , among others , the development of new systems of literature reviews , machine - searchable archives for data and publications , and virtual observatories .
as customary in these reports , we highlight some of the research topics that have seen particular interest over the most recent triennium , specifically active - region magnetic fields , coronal thermal structure , coronal seismology , flares and eruptions , and the variability of solar activity on long time scales .
we close with a collection of developments , discoveries , and surprises that illustrate the range and dynamics of the discipline . * * = division ii + commission 10 solar activity + + = president carolus j. schrijver + vice - president lyndsay fletcher + past president lidia van driel - gesztelyi + organizing committee ayumi asai , paul s. cally , + paul charbonneau , sarah e. gibson , + daniel gomez , siraj s. hasan , + astrid m. veronig , yihua yan + .overview of commission 10 leadership and triennial reports ( as available in ads ) from 1961 onward , i.e. for the period that c10 operated under the banner of `` solar activity '' or `` activit solaire '' .
reports flagged with an asterisk appear not to be available on line . [ cols="<,<,<",options="header " , ] the community exchanges information efficiently at scientific meetings .
fig .
[ fig : meetings ] shows that the number of such events tends to increase over the years , with marked fluctuations from year to year , averaging around 40 meetings per year over the past decade .
the symposia supported by the iau through commission 10 in the last three years are listed in table [ tab : meetings ] .
summer schools in which new generations of researchers are given broader or deeper views into the community s activities appear to grow slowly in frequency , trending towards about five events per year ( dashed line in fig .
[ fig : meetings ] ) .
although the scientific community working on `` solar activity '' appears healthy and growing , there is a clear need to improve how we communicate the excitement about our science to the general public : for example , only 8 in the most recent 400 press releases and news articles listed by the aas were related to some aspect of solar activity . |
at the center of the milky way lurks a supermassive black hole ( smbh ) , whose presence and mass , @xmath34@xmath4 , has been ferreted out through its strong gravitational field ( e.g. , ghez et al . 2003a , b ; schdel et al .
2002 , 2003 ) , but whose associated radiative emission has remained more elusive . for decades
it was only detected at radio wavelengths ( see , e.g. , melia & falcke 2001 ) , where it appears to show low amplitude intensity variations with a quasi - periodicity of @xmath3 106 d ( zhao , bower , & goss 2001 ) .
recent observations detected it at x - ray wavelengths , where it appears to be composed of the following two distinct components ( baganoff et al .
2001 , 2003a , b ; goldwurm et al .
2003 ; porquet et al .
2003 ) : ( 1 ) a steady state , which has been stable to within @xmath310% over the past 4 years and which is spatially resolved with a size that corresponds to the bondi radius ( @xmath31@xmath5 ) , and ( 2 ) an unresolved variable component , which has a flux density that rises above the quiescent level by an order of magnitude for @xmath3 1 hour approximately once a day .
these detections at opposite ends of the electromagnetic spectrum , along with existing infrared ( ir ) limits ( e.g. , serabyn et al .
1997 ; morris et al .
2001 ; hornstein et al .
2002 ) reveal that overall , sgr a * is remarkably weak , with a bolometric luminosity of only 10@xmath6 ergs s@xmath7 or , equivalently , @xmath8 .
detection of emission at ir wavelengths would be a very powerful constraint for the proposed models of the accretion flow onto the smbh as well as possible outflows ( e.g. , markoff et al .
2001 ; liu & melia 2002 ; yuan et al .
2003a ) . at mid - ir wavelengths ( @xmath98 @xmath0 m ) ,
the detectability of a source is limited by thermal emission from dust ( stolovy et al . 1996 ; morris et al . 2001 ) . at near - ir wavelengths ( 1 - 2 @xmath0 m ) , detection of an associated point source
is difficult due to the confusion with stellar sources ( e.g. , close et al .
1995 ; genzel et al .
1997 ; hornstein et al .
2002 ) . in the 3 - 5 @xmath0 m window ,
the stellar and dust emission are both declining , potentially allowing an unambiguous detection of a radiative counterpart at a wavelength intermediate to the radio and x - ray regimes . in this paper , we present new
l(3.8 @xmath0 m ) images from the w. m. keck ii 10-meter telescope of the galactic center showing a near - ir counterpart to sgr a*. section 2 describes the observations and 3 summarizes the data analysis .
section 4 reports the detection of a new l source and its properties . finally ,
5 discusses possible emission mechanisms for the new source and concludes that it is indeed associated with the smbh .
adaptive optics l ( @xmath10 = 3.8 @xmath0 m & @xmath11 = 0.7 @xmath0 m ) bandpass observations of the galactic center were obtained with the w. m. keck ii 10 m telescope using the facility near - infrared camera , nirc2 ( matthews et al .
, in prep ) on the following 4 nights : 2002 may 31 , 2003 june 10 , 16 , & 17 ( ut ) . the positions of irs 16ne , nw , & sw with respect to irs 16c in the final map ( see 3 ) compared to that reported in ghez et al .
( 2003b ) established a plate scale of 9.93 @xmath12 0.05 mas pix@xmath7 , corresponding to a field of view of 102 @xmath13 102 for nirc2 s narrow field camera .
using an r=13.2 mag natural guide star located 30@xmath5 from sgr a * , we were able to achieve a resolution as high as 80 mas , which corresponds to the diffraction limit , and strehl ratios as high as @xmath30.4 .
each night , several images , composed of 0.2 to 0.5 sec coadded exposures with effective exposure on the sky of 20 to 60 sec , were collected using a 5 position dither pattern with offsets of @xmath31 arcsec .
the standard image reduction steps of sky subtraction , flat fielding , and bad pixel removal were carried out on each image . visual inspection of the reduced individual images eliminated those that were degraded by poor adaptive optics tip - tilt corrections , which occasionally generated double peaked psfs .
inclusion in the final map for the remaining images was based on their estimated strehl ratios , which were required to be @xmath90.2 .
all of the selected images were registered and averaged together using the centroid of irs 16c s light distribution .
sources were identified and their positions and relative intensities were quantified using the point spread function ( psf ) fitting program starfinder ( diolaiti et al . 2000 ) .
three bright point sources ( irs 16c , 16nw , & 33n ) were entered into starfinder s algorithm for generating a model psf .
uncertainties in this process were established by dividing the data set up into independent maps composed of 2 to 3 frames , re - running this procedure on each map , and taking root - mean - square values for the astrometric and relative photometric values .
this also allowed an examination of the data set for shorter timescale flux density variations .
measurements reported here are photometrically calibrated based on absolute l photometry for several bright stars within our fied of view obtained by simons & becklin ( 1996 ) . from the 8 stars in common , we selected the following 4 stars , which are not known to be variable or spatially extended at l or at k ( 2.2 @xmath0 m ) , as calibration sources : irs 16ne ( 7.77 mag ) , irs 16sw - e ( 8.30 mag ) , irs 16nw ( 8.87 mag ) , & irs 16c ( also labeled irs 16c - w ; 8.91 mag ) .
our multiple measurements of these sources reveal no evidence for variability , confirming their suitability as local standards .
we estimate the uncertainty in the apparent magnitude calibration for each image to be 0.1 mag from the standard deviation of the mean of the calibration sources .
all the reported measurements are based on a zero point of 248 jy for the l magnitude scale ( tokunaga 2000 ) .
the inferred position of the smbh in the l maps is based on its dynamically determined position . using the orbital solutions for s0 - 1 , s0 - 2 ,
s0 - 3 and s0 - 4 reported in ghez et al .
( 2003b ) , we localize the smbh in each of the l images to within @xmath12 5 - 10 mas , where these uncertainties are limited by inaccuracies in the centroids of the stars in the l maps .
figure [ fig_lp ] shows a 12 @xmath13 12 region of the combined images centered on the smbh s location for each of the four nights reported here . in 2003
, a new source is detected . with the exception of this new source all other detected l sources in figure [ fig_lp ] are coincident with stars detected at 2.2 @xmath0 m ( ghez et al .
the new source distinguishes itself in several ways .
it is an unresolved point source that has a negligible average offset of 12 @xmath12 6 mas from the smbh .
in contrast to stars in its vicinity , the new l source is remarkably variable .
this is shown in figure [ fig_phot ] , which plots the new source s dereddened flux density along with that of all sources that are nearby ( r @xmath14 05 ) , are comparably bright ( l @xmath15 13.6 mag ) , and do not suffer from confusion ; all the sources are dereddened assuming an l extinction of 1.83 mag , which is based on a visual extinction of 30 mag from rieke , rieke , & paul ( 1989 ) and an extinction law from moneti et al .
on 2003 june 10 the new source was at its brightest at 12.28 mag and then it dimmed by a factor of 3 ( 13 @xmath1 ) to 13.38 mag by 2003 june 17 , with a factor of 2 ( 12 @xmath1 ) change occurring between june 16 and 17 .
subdivision of the data into shorter time blocks of 60 second exposures , which correspond to elapsed time of 100 - 250 seconds depending on how images were selected ( see 3 ) , shows significant substructure ( see figure [ fig_phot ] ) ; on 2003 june 17 the new source is observed to have dimmed by a factor of 2 , with a significance of 5 @xmath1 .
this also increases the range of intensity variations observed over a week to a factor of 4 .
table [ tab_sum ] provides a summary of this source s measured properties , which includes its position with respect to the smbh as well as its apparent magnitude and dereddened flux density at 3.8 @xmath0 m . in our 2002 map ,
the new l source is blended with the star s0 - 2 , which at this time was experiencing its closest approach to the smbh with a projected separation of a mere 14 mas 2003a , b ) .
this confused the identification of the l source in this first epoch map , as well as those obtained at the vlt on 2002 august 29 ( clnet et al .
2003 , genzel et al .
2003 ) = l@xmath16 + 0.82 mag & l@xmath17 = l@xmath18 + 0.37 mag .
this is primarily a consequence of a different choice of calibration sources , which includes known variable and extended stars in their sample ( see 3 ) ; clnet et al .
( 2003 ) use irs 16c , 29n , 16cc , 21 , 33se , & mpe+1.6 - 6.8 and genzel et al .
( 2003 ) restrict their analysis of the same data set to irs 16cc & 33n . ] in our 2003 maps , s0 - 2 is 87 mas away from the dynamical center and is clearly separated from the newly identified l source ( see figure [ fig_lp ] ) . since s0 - 2 shows no evidence for significant variability in the k - band , where there is negligible confusion with other sources ( ghez et al .
2003b ) , we use the l mag of s0 - 2 measured in 2003 june 16 & 17 , when the contrast between s0 - 2 and the new source is smallest , ( @xmath14l@xmath19 = 13.07 ) to subtract the s0 - 2 contamination from the new source in 2002 may 31 .
this results in an inferred brightness of l=13.2 mag , which is close to the average of the values measured in 2003 june .
a comparison of the 2002 position with the average 2003 position limits its proper motion to 400 @xmath12 300 km s@xmath7 .
the newly identified l source has several properties that indicate that it is affiliated with the smbh .
first , its location on the plane of the sky is coincident with the smbh to within 18 mas ( 1 @xmath1 ) .
second , it appears to be a stationary source with an upper limit on its transverse motion of 700 km s@xmath7 , which differs from the proper motions of @xmath206,000 km s@xmath7 for other sources that have come within 10 mas of the dynamical center ( ghez et al .
2003b ; schdel et al .
third , it is a variable source with significant flux density changes observed on timescales as short as 40 min .
fourth , it is an extremely red object . in the k - band ,
a stationary source consistent with the location of the dynamical center would have to be fainter than at least k @xmath3 15.5 mag to be consistent with the observations of ghez et al .
( 2003b ) and hornstein et al .
the k - l color of the new source must therefore be greater than 2.1 mag , which is at least 1.0 mag redder than all the nearby stars used as comparison stars in figure [ fig_phot ] .
such a red color is expected for sgr a * from the sub - mm detections ( e.g. , falcke et al .
1998 ; zhao et al .
2003 ) and 2.2 @xmath21 limits ( hornstein et al . 2002 ) shown in figure 3 .
we conclude that the new l source is associated with the smbh or , equivalently , sgr a * , and therefore refer to it as sgra*-ir .
several mechanisms associated with a smbh can give rise to infrared variability , including gravitational lensing ( e.g. , alexander & loeb 2001 ) , disk illumination ( cuadra , nayakshin , & sunyaev 2003 ) , star - disk collisions ( nayakshin & sunyaev 2003 ) , and physical processes in the smbh s accretion flow ( e.g. , markoff et al .
2001 ; liu & melia 2002 ; yuan , quataert , & narayan 2003a , b ) . assuming that we are witnessing a single phenomenon , the observed short timescales for the variations , @xmath340 min , rule out the first two possibilities and the longer timescales eliminate the third .
we therefore suggest that the detected 3.8 @xmath0 m emission emanates from either an accretion flow or an outflow quite close to the smbh , within 5 au , or equivalently 80 r@xmath2 . while the shortest timescale variations observed at 3.8 @xmath0 m are comparable to the x - ray flare timescales ( baganoff et al .
2001 , 2003b ; goldwurm et al . 2003 ; porquet et al .
2003 ) , the probability that such an x - ray flare occurred during any of these observations is less than 0.05 and the probability that such x - ray flares occurred or were in progress during all of our observations is less than 10@xmath22 . nonetheless , the similarity in the timescales for variation at 3.8 @xmath0 m and at x - ray wavelengths lead us to suggest that the variable 3.8 @xmath0 m emission is produced by a mechanism related to that which produces the x - ray flares .
we presume any associated variable x - ray emission during our observations was below the quiescent level .
several models of the flared x - ray emission have been proposed , invoking physical processes such as elevated accretion rates or enhanced electron acceleration from mhd turbulence , reconnection , or weak shocks and emission mechanisms including bremsstrahlung , synchrotron , and inverse compton ( markoff et al . 2001 ; liu & melia 2001 , 2002 ; yuan et al .
2003a , b ) .
models that account for the low variability amplitude at radio frequencies generally invoke acceleration of some fraction of the electrons into the power - law tail of the electron energy distribution . in these models ,
the x - ray flares arise from either synchrotron emission , if a small fraction of the electrons are energetic enough ( @xmath23 ) , or by synchrotron - self compton ( ssc ) upscattering of sub - mm and infrared photons , for a larger but less dramatically accelerated population of electrons . in both cases ,
a significant amount of 3.8 @xmath0 m emission can be produced from the direct synchrotron emission of the accelerated population of electrons ( see figure [ fig_sed ] ) and the variability may be induced by small changes in the fraction of electrons in the high - energy power - law tail of the electron energy distribution or by changes in the value of the power - law ( yuan et al .
for the two cases that produce 3.8 @xmath0 m emission via synchrotron emission , the ssc model predicts correlated ir and x - ray flares , while the synchrotron only model can more readily allow for an ir only flare , depending on the spectrum of the accelerated electron energy distribution .
it therefore appears at present that the synchrotron only model best fits the available data ( e.g. , yuan et al . 2003b ) .
furthermore , in contrast to the x - ray flares , which suggest that processes giving rise to a high - energy power - law tail in the electron energy distribution occur infrequently ( @xmath92% of the time ) , the persistently detected variable 3.8 @xmath0 m emission suggests that such processes are in effect much more often .
more generally , the 3.8 @xmath0 m emission opens up a continuously accessible window for studying the conditions and physical processes within the plasma that is either accreting onto or outflowing from the central black hole .
the authors thank j. aycock , r. campbell , g. hill , c. sorensen , m. van dam , & c. wilburn at the keck observatory for their help in obtaining the observations & g. duchne , j. lu , f. melia , r. narayan , f. yuan , & an anonymous referee for useful comments .
support for this work was provided by nsf grant ast-9988397 & the nsf science & technology center for ao , managed by ucsc ( ast-9876783 ) .
dt & bts are supported by the sirtf project .
sirtf is carried out at the jpl , operated by the caltech , under contract with nasa .
the keck observatory is operated as a scientific partnership among caltech , the university of california , & nasa .
the observatory was made possible by the generous financial support of the keck foundation .
ghez , a. m. , duchne , g. , matthews , k. , hornstein , s. d. , tanner , a. , larkin , j. , morris , m. , becklin , e. e. , salim , s. , kremenek , t. , thompson , d. , soifer , b.t .
, neugebauer , g. , mclean , i. 2003a , , 586 , l127 lllrrr 2002 may 31 & average & 13.23 @xmath12 0.10 & 6.9 @xmath12 0.7 & 0 @xmath12 @xmath246 & -7 @xmath12 @xmath247 + & 11:29 & 12.78 @xmath12 0.18 & 10.35 @xmath12 1.9 & & + & 11:35 & 12.97 @xmath12 0.20 & 8.7 @xmath12 1.7 & & + & 11:41 & 13.01 @xmath12 0.20 & 8.4 @xmath12 1.7 & & + & 11:47 & 13.64 @xmath12 0.27 & 4.7 @xmath12 1.3 & & + & 11:52 & 13.47 @xmath12 0.25 & 5.5 @xmath12 1.4 & & + 2003 jun 10 & average & 12.28 @xmath12 0.05 & 16.4 @xmath12 0.8 & -8 @xmath12 @xmath249 & -1 @xmath12 10 + & 12:06 & 12.22 @xmath12 0.07 & 17.4 @xmath12 1.2 & & + & 12:10 & 12.32 @xmath12 0.07 & 15.8 @xmath12 1.1 & & + 2003 jun 16 & average & 12.54 @xmath12 0.05 & 12.9 @xmath12 0.6 & -15 @xmath12 @xmath249 & 9 @xmath12 @xmath248 + & 11:20 & 12.56 @xmath12 0.08 & 12.6 @xmath12 1.0 & & + & 11:28 & 12.45 @xmath12 0.08 & 14.0 @xmath12 1.1 & & + & 11:38 & 12.80 @xmath12 0.09 & 10.2 @xmath12 0.9 & & + 2003 jun 17 & average & 13.38 @xmath12 0.04 & 5.9 @xmath12 0.2 & -12 @xmath12 13 & -7 @xmath12 18 + & 10:32 & 13.25 @xmath12 0.09 & 6.7 @xmath12 0.6 & & + & 10:36 & 13.24 @xmath12 0.09 & 6.7 @xmath12 0.6 & & + & 10:41 & 13.10 @xmath12 0.09 & 7.7 @xmath12 0.7 & & + & 10:45 & 13.01 @xmath12 0.08 & 8.4 @xmath12 0.7 & & + & 10:49 & 13.22 @xmath12 0.09 & 6.9 @xmath12 0.6 & & + & 10:59 & 13.40 @xmath12 0.10 & 5.8 @xmath12 0.6 & & + & 11:16 & 13.72 @xmath12 0.11 & 4.3 @xmath12 0.5 & & + & 11:24 & 13.84 @xmath12 0.12 & 3.9 @xmath12 0.5 & & + | we report the detection of a variable point source , imaged at l(3.8 @xmath0 m ) with the keck ii 10 m telescope s adaptive optics system , that is coincident to within 18 mas ( 1 @xmath1 ) of the galaxy s central supermassive black hole and the unique radio source sgr a*. while in 2002 this source ( sgra*-ir ) was confused with the stellar source s0 - 2 , in 2003 these two sources are separated by 87 mas allowing the new source s properties to be determined directly . on four separate nights ,
its observed l magnitude ranges from 12.2 to 13.8 , which corresponds to a dereddened flux density of 4 - 17 mjy ; no other source in this region shows such large variations in flux density - a factor of 4 over a week and a factor of 2 over 40 min . in addition , it has a k - l color greater than 2.1 , which is at least 1 mag redder than any other source detected at l in its vicinity .
based on this source s coincidence with the galaxy s dynamical center , its lack of motion , its variability , and its red color , we conclude that it is associated with the central supermassive black hole .
the short timescale for the 3.8 @xmath0 m flux density variations implies that the emission arises quite close to the black hole , within 5 au , or 80 r@xmath2 .
we suggest that both the variable 3.8 @xmath0 m emission and the x - ray flares arise from the same underlying physical process , possibly the acceleration of a small populations of electrons to ultrarelativistic energies .
in contrast to the x - ray flares which are only detectable @xmath32% of the time , the 3.8 @xmath0 m emission provides a new , constantly accessible , window into the physical conditions of the plasma in close proximity to the central black hole . |
from the figures ( [ fig2 ] , [ fig3 ] ) one can see that in the inclusive setup of the process of charged pairs production by interaction of linearly polarized high energy photon with proton distribution of recoil proton has rather essential azimuthal asymmetry , from 0.02 at the relatively small polar angles @xmath19 up to @xmath70 at @xmath71 . for the cases of @xmath67 pair , @xmath68 pair and @xmath69 pair production and also @xmath72 as function of scattering angle @xmath19 . ] in exclusive setup for processes with more heavy particles than e@xmath73e@xmath74 , mentioned asymmetry increases .
particularly interesting is the process of @xmath69 pair photoproduction .
one can see that azimuthal asymmetry of recoil proton in this process reaches the value @xmath75 at the region of small transferred momentum or for polar angles close to the value @xmath76 .
this features of @xmath69 pair photoproduction process allows one to hope that this process can be considered as the polarimetric process .
we shall discuss this process from this point of view in the more details in the next work .
the inclusive on recoil proton distribution is the sum on all possible channels including fermion ( @xmath67 , @xmath68 , @xmath77 ) and pseudoscalar meson ( @xmath69 , @xmath78 ) pairs .
production of heavy resonances such as @xmath79 meson can be excluded using experimental cuts .
the suggested method of measuring the recoil distributions can provide the independent way to control the luminosity and polarization properties of photon beam . in paper @xcite the photoproduction of electron
positron pair on electron was considered in lowest order of pt .
the radiative corrections to cross section were considered in paper @xcite and in all orders of pt on parameter @xmath80 in paper @xcite both for unpolarized case .
it turns that for @xmath81 our results can be applied for photoproduction on nuclei with relevant modification of @xmath63 .
the radiative corrections can change the values @xmath82 , @xmath83 in frames of 12@xmath84 .
the proton recoil momentum measurements can as well be arranged in @xmath85 and @xmath86 experiments with initial proton at rest .
using the weizscker
williams approximation the corresponding cross sections can be written as @xmath87\dd\sigma^{\gamma { \mathrm { p}}\to a\bar{a}{\mathrm { p'}}}\end{gathered}\ ] ] for electron proton collisions and @xmath88 \dd\sigma^{\gamma { \mathrm { p}}\to a\bar{a}{\mathrm { p'}}},\end{gathered}\ ] ] for proton proton collisions with @xmath89 , @xmath90 is the energy of initial electron or proton and @xmath91 are given above . inferring these formulae we supposed that the transversal momentum of projectile ( e or p ) does not exceed @xmath92 .
the polarization vector of virtual photon interacting with proton at rest is directed along this projectile transverse momentum .
one of us ( eak ) is grateful to institute of physics , sas .
we are grateful to participants of seminar of theoretical physics sas .
one of us ( y.p ) is grateful to grant stsu-3239 .
the work was partly supported also by slovak grant agency for sciences vega , grant no . 2/4099/26 .
more exact formula which takes into account power corrections for recoil proton momentum has a form @xcite @xmath93 under condition @xmath94 upper branch of ( [ yupcom1 ] ) passes to @xmath95 with @xmath96 invariant mass squared of pair produced
. 99 d. benaksas and r. j. morrison , phys . rev .
* 160 * ( 1967 ) 1245 .
v. n. baier , e. a. kuraev , v. s. fadin and v. a. khoze , phys .
* 78 * , 293 ( 1981 ) .
a. i. akhiezer , m. p. rekalo , _ electrodynamics of hadrons _
( in russian ) , naukova dumka , kiev , 1977 . e. a. vinokurov and e. a. kuraev , zh . eksp
. teor .
fiz . * 63 * , 1142 ( 1972 ) .
e. a. vinokurov , e. a. kuraev and n. p. merenkov , zh . eksp .
* 66 * , 1916 ( 1974 ) .
d. ivanov and k. melnikov , phys .
d * 57 * , 4025 ( 1998 ) v. f. boldyshev , e. a. vinokurov , n. p. merenkov and yu .
p. peresunko , phys .
part . nucl . * 25 * , 292 ( 1994 ) . | for high energy linearly polarized photon
proton scattering we have calculated the azimuthal and polar angle distributions in inclusive on recoil proton experimental setup .
we have taken into account the production of lepton and pseudoscalar meson charged pairs .
the typical values of cross sections are of order of hundreds of picobarn .
the size of polarization effects are of order of several percents .
the results are generalized for the case of electroproduction processes on the proton at rest and for high energy proton production process on resting proton .
we have considered below the experimental setup of processes of charged pair @xmath0 production ( pseudoscalars , leptons ) by high energy photon scattering on proton at rest frame with following detection of recoil proton @xmath1 two different mechanisms of pair production must be considered .
one corresponds to the pair creation by two photons bethe heitler ( bh ) .
another one is the bremsstrahlung ( b ) , which corresponds to the case when pair is created by single virtual photon ( we have implied the lowest in qed coupling constant @xmath2 contributions ) .
the contribution of b type is suppressed compared with one of bh type by factor @xmath3 . as for interference of b and bh
amplitudes it is exactly zero in the inclusive on recoil proton setup we have considered below .
the accuracy of formulae given below are determined by the terms we have omitted systematically compared with terms of order of unity @xmath4 in the peripheric kinematical region @xmath5 effectively works the infinite momentum frame ( imf ) or sudakov @xcite parametrization of transferred momentum and the 4momenta of final particles @xmath6 from the on mass shell condition of recoil proton @xmath7 , one infers @xmath8 we use here the smallness of @xmath9 compared with @xmath10 .
here @xmath11 invariant mass square of pair , assumed to be of order @xmath12 .
for the case of large @xmath13 one can put considered @xmath14 to @xmath15 .
the ratio of transversal and longitudinal component of momentum of recoil proton ( laboratory frame implied ) is @xmath16 this relation , first mentioned in paper of benaksas and morrison @xcite , can be written in different form in terms of the value for 3vector of momentum of recoil proton @xmath17 @xmath18 with @xmath19 is the angle between the directions of initial photon and recoil proton in laboratory frame ( see more exact formula in appendix a )
. matrix element of charged lepton or pion pair production in lowest order of qed perturbation theory ( keeping in mind bh mechanism ) has the form @xmath20 with proton current defined as @xmath21}{4m}f_2(q^2)]u(p),\ ] ] and @xmath22 proton form factors .
compton lepton tensor has the form @xmath23v(q_+),\ ] ] similarly compton pion tensor @xmath24 these tensors obey the gauge invariance requirements @xmath25 .
using gribov prescription for green function of virtual photon in feynman gauge and omitting small contributions in frames of declared accuracy @xmath26\approx \frac{2}{s}\tilde{p}^\mu k^\nu,\ ] ] one can put the matrix element , extracting explicitly the factor @xmath27 in form @xmath28 both light cone projections of proton current and compton tensors are finite in large @xmath27 limit .
summing on proton spin states one has for proton current projection square @xmath29.\ ] ] expressing the phase volume of the final particles in terms of sudakov variables @xcite @xmath30 where we introduce the unit factor @xmath31 and besides we have used @xmath32 further operations , summing over spin states of leptons of square of matrix element , performing the integration over pair energy fractions @xmath33 , @xmath34 , ( @xmath35 ) and its transversal momentum @xmath36 , ( conservation low provides @xmath37 ) , and using the photon polarization matrix @xmath38_{ij}$ ] , ( @xmath39 stokes parameters , @xmath40 unite matrix , @xmath41 pauli matrices ) is straightforward but tedious . the result can be written in the form @xmath42 where @xmath43 .
the azimuthal angle @xmath44 is the angle between two transversal to photon direction vectors : photon linear polarization @xmath45 and @xmath46 ; @xmath47 is angle between @xmath45 and axes @xmath48 ; @xmath49 is degree of photon linear polarization and @xmath50 @xmath51 is azimuthal asymmetry @xmath52 in the equations ( [ eq11 ] , [ eq12 ] ) quantities @xmath53 , @xmath54 are @xmath55 it is interesting to consider distribution @xmath56 of recoil proton on the value @xmath57
. calculations of this distribution were carried out on the base of formula , which is obtained from ( 10 , 11 ) after substitution @xmath58 ( see ( [ eq6 ] ) ) @xmath59 at the fig .
[ fig1 ] the distributions @xmath60 for each of considered processes are depicted .
for numerical calculation we used the dipole approximation @xcite @xmath61}{0.71 ^ 2}\big)^2 } , \\
f_e = f_1-f_2\frac{q^2}{4m^2},\quad f_m = f_1+f_2.\end{gathered}\ ] ] with @xmath62 - anomalous magnetic moment of proton .
function @xmath63 in dipole approximation has the form @xmath64}{(0,71)^2}+1)^4}.\ ] ] at the fig .
[ fig2 ] the asymmetries @xmath51 as function of momentum @xmath57 for each of considered processes are shown . at the fig .
[ fig3 ] mentioned asymmetries as function of the scattering angle @xmath19 are shown .
the ratio @xmath65 can be considered as averaged over all processes asymmetry which estimate the total influence of initial photon linear polarization on the value of recoil proton azimuthal asymmetry .
this value is also presented at the fig .
[ fig3 ] . in units of pbn mev@xmath66 for the cases of @xmath67 pair , @xmath68 pair and @xmath69 pair production as function of @xmath57 . ] for the cases of @xmath67 pair , @xmath68 pair and @xmath69 pair production as function of @xmath57 . ] |
the study of the mercury - oxide family hgba@xmath0ca@xmath13cu@xmath14o@xmath15 is of particular interest since these cuprates possess the highest transition temperatures @xmath4 of all known high - temperature superconductors ( htsc ) . for the hg - compound containing one cuo@xmath0-layer , hgba@xmath0cuo@xmath5 ( hg-1201 ) , @xmath4 reaches values as high as 98 k at ambient pressure @xcite and
can reportedly be varied over the entire range from extreme under- to overdoped solely by changing the oxygen content @xcite .
the hg - compound containing three cuo@xmath0-layers exhibits the highest ambient - pressure transition temperature ( @xmath6 134 k ) reported to date , increasing to @xmath16 164 k under 30 gpa pressure @xcite .
high - pressure experiments are potentially an important tool to further our understanding of the superconducting state @xcite . for the hg - compounds the dependence of @xmath4 on pressure was first studied by klehe et al .
@xcite on optimally doped hg-1201 where @xmath17 was found to increase linearly with hydrostatic pressure to 1 gpa at the rate + 1.75 k / gpa .
an identical initial pressure dependence was subsequently observed by klehe et al .
@xcite for the two- and three - layer compounds hgba@xmath0cacu@xmath0o@xmath3 ( hg-1212 ) and hgba@xmath18ca@xmath0cu@xmath1o@xmath19 ( hg-1223 ) ; this parallelism in the @xmath17(p)-dependence for all three optimally doped hg - compounds was later confirmed to much higher pressures ( 30 gpa ) by gao et al .
@xcite who observed that @xmath4 initially increases , but passes through a maximum and decreases at higher pressure .
in fact , for the majority of htsc compounds near optimal doping , @xmath4(p ) is found to follow a similar bell - shaped curve @xcite .
cao et al .
@xcite studied the pressure dependence of @xmath4 for hg-1201 to 1.5 gpa at widely differing doping levels by varying the oxygen concentration ; qiu et al .
@xcite extended these studies to 18 gpa for selected samples with @xmath4 values at or above 80 k. jansen and block @xcite have attempted a quantitative analysis of these experiments using an indirect - exchange pairing mechanism between conduction electrons via oxygen anions ; the doped oxygen thus fulfills a dual role as both hole dopant and cooper pair mediator . in this model calculation pressure effects on @xmath4
are assumed to originate solely from a change in unit cell volume ; possible rearrangements of oxygen defects within the unit cell when pressure is applied are not considered . before attempting to gain information about the superconducting state from a detailed analysis of @xmath4(p)-data on the superconducting oxides , it is essential to know , on a microscopic scale , what structural changes actually occur under pressure .
it would be hoped that @xmath20 is indeed a reversible , single - valued function of pressure , as in conventional superconductors .
this has appeared to be the case in the vast majority of experiments on the superconducting oxides where the pressure was changed at ambient temperature .
surprisingly , experiments carried out by sieburger et al .
@xcite on tl@xmath0ba@xmath0cuo@xmath3 ( tl-2201 ) revealed that the initial dependence of @xmath4 on pressure can change by an order of magnitude , depending on the temperature at which the pressure is varied ; even the sign of @xmath21 may change .
@xmath4(p ) is thus _ not _ a unique function of pressure but depends strongly on the detailed temperature - pressure history of the sample .
because this anomalous behavior is a sensitive function of the concentration of interstitial oxygen @xmath22 in tl-2201 , it has been attributed @xcite to relaxation effects within the oxygen sublattice ; indeed , evidence for strong relaxation effects in @xmath4 from oxygen ordering were first obtained by veal et al .
@xcite in temperature - quench experiments on strongly underdoped yba@xmath0cu@xmath1o@xmath2 ( y-123 ) .
evidence for relaxation effects in y-123 from high - pressure data was obtained somewhat later @xcite . in the meantime
many oxide superconductors have been found which exhibit pressure - induced relaxation effects @xcite , including tlsr@xmath0cacu@xmath0o@xmath23 ( tl-1212 ) @xcite , nd-123 @xcite , gd-123 @xcite , superoxygenated la@xmath0cuo@xmath24 @xcite , and possibly sr@xmath0cuo@xmath0f@xmath25 @xcite .
for all these systems , except tl-2201 , relaxation effects become activated for temperatures above 200 k. in tl-2201 two distinct relaxation processes are activated in the temperature ranges 15 - 100 k and 200 - 300 k , respectively @xcite . in view of the importance of relaxation phenomena in many high - t@xmath26 superconductors
, it would be of interest , before carrying out detailed comparisons between theory and experiment , to determine whether or not they play a role in the hg - oxide system where the values of @xmath20 are relatively high , and , at least in hg-1201 , the oxygen defect structure is known to be quite complex @xcite .
in hg-1212 and hg-1223 evidence for oxygen ordering along the c - axis has been obtained in energy dispersive synchrotron x - ray diffraction experiments for pressures above @xmath27 15 gpa @xcite ; the effect of this ordering on @xmath4 was not studied . in this paper
we show evidence for the existence of weak pressure - induced relaxation effects in @xmath4 for both under- and overdoped hg-1201 which are activated at temperatures above 200 k. previous @xmath20 studies by us on optimally doped hg-1201 @xcite gave no evidence for relaxation behavior . in the overdoped hg-1201 samples the pressure dependence @xmath20 to 0.8 gpa is clearly nonlinear , providing evidence for a possible pressure - induced electronic and/or structural modification .
in the present experiments one underdoped sample and two overdoped samples ( a and b ) of polycrystalline hgba@xmath0cuo@xmath5 with transition temperatures in the range 40 - 50 k were studied .
typical sample sizes were 2@xmath281@xmath281 mm@xmath29 with masses ranging from 4 to 12 mg .
all samples were synthesized in an evacuated quartz tube as described previously @xcite .
both the single underdoped sample and the overdoped sample a were obtained from a single synthesis by subsequent annealing to produce samples with significantly different oxygen contents but similar values of @xmath4 . the overdoped sample a ( @xmath4(1 bar ) @xmath30 47.81 k )
was annealed in 150 atm .
oxygen at 220@xmath31 for 100 h followed by a slow cool to room temperature , whereas the underdoped sample ( @xmath4(1 bar )
@xmath8 47.55 k ) was annealed in flowing argon at 400@xmath32 for 24 h and then rapidly cooled to room temperature . the overdoped sample b ( @xmath4(1 bar )
@xmath8 40.0 k ) was obtained from a second synthesis in powder form and annealed in 30 atm .
oxygen at 200@xmath32 for 100 h. neutron diffraction experiments on these samples determined the oxygen contents at x @xmath33 0.22 @xcite and 0.12 @xcite for the overdoped sample a and the underdoped sample , respectively .
the he - gas pressure system ( harwood ) used in this study is capable of generating hydrostatic pressures to 1.4 gpa .
the cube pressure cell ( unipress ) is inserted into a two - stage closed - cycle refrigerator ( leybold ) which operates in the temperature range 2 - 320 k. the pressure in the cell can be changed at any temperature above the melting curve t@xmath34 of the he pressure medium @xcite ( for example , @xmath35 13.8 k at 0.1 gpa and @xmath35 52.4 k at 0.8 gpa ) . the decrease in pressure upon cooling
is kept to below 20% by the introduction of a sizeable room temperature `` dead volume '' .
unless otherwise stated , the value of the pressure given was determined at a temperature near @xmath4 .
the superconducting transition of the hg-1201 samples is measured by the ac susceptibility technique using a miniature primary / secondary coil system located inside the 7 mm i.d .
bore of the pressure cell . in the present studies a magnetic field of 1.1 oe ( rms ) at 507 hz
is applied along the long dimension of the sample .
further details of the experimental setup are given elsewhere @xcite .
1 shows the transition to superconductivity for the overdoped hg-1201 sample a at ambient pressure .
the transition is measured by sweeping through the temperature range 10 k to 55 k over a 2 h period , yielding the value @xmath4(1 bar ) @xmath8 47.81 k ( inset in fig .
1 ) for the superconducting onset .
the size of the transition corresponds to full shielding , if the demagnetization factor @xmath36 is neglected @xcite .
the transition temperature in the subsequent measurements under high pressure is determined by shifting the upper 15% of the full transition along the temperature axis until overlap is achieved with the initial measurement .
only the upper portion of the transition is used because this region is independent of the applied magnetic field . through this procedure
an accuracy of @xmath11 20 mk for the relative @xmath4 shifts under pressure could be achieved . in most experiments
the transition was remeasured for verification .
the first set of high - pressure data for the overdoped hg-1201 sample a ( open circles in fig .
2a ) was taken by applying 0.16 gpa at 298 k and holding this pressure for one hour , followed by a rapid cooldown ( 3 - 4 k / min ) to measure @xmath4 . at temperatures near @xmath4 the measured pressure had decreased to 0.12 gpa due to the thermal contraction of the helium pressure medium upon cooling .
the sample was then heated back to 298 k , the pressure increased and @xmath4 remeasured .
this procedure was repeated up to the highest pressure of 0.61 gpa ( 0.54 gpa at 50 k ) . the pressure dependence @xmath20
is seen to be markedly nonlinear , an anomalous result in this pressure range for high - t@xmath26 oxides .
the second set of data on this sample ( solid squares in fig .
2a ) was taken by applying 0.79 gpa at 298 k
( 0.71 gpa at 50 k ) followed by a one hour anneal at this temperature . after measuring @xmath4 and heating back up to 298 k
, the pressure was decreased by @xmath27 20% , held for one hour at 298 k , and @xmath4 remeasured .
this procedure was then repeated several times until the pressure was fully released ; @xmath4 is seen to return to its initial value at 47.8 k. for pressure changes at room temperature ( rt ) , @xmath37 thus appears to be a reversible , but markedly nonlinear , function of pressure with initial slope @xmath38
-1.55 @xmath11 0.26 k / gpa .
we now determine the hydrostatic pressure dependence of @xmath4 for pressure changes at low temperatures , an experiment that can best be carried out using a he - gas high - pressure system . in the third set of data on the overdoped sample a ( solid diamonds in fig .
2a ) , we first apply and hold 0.76 gpa pressure at 298 k for 10 min before cooling down rapidly to measure @xmath17 ( @xmath39 0.69 gpa at 50 k ) . some pressure was then added at 60 k to obtain 0.75 gpa and @xmath4 was remeasured .
the pressure was then successively decreased at low temperatures ( t 55 k ) and @xmath40 redetermined .
the value of @xmath4 at ambient pressure is now seen to lie 0.26 k below its initial value . only after annealing the sample at 298 k for 40
min does @xmath4 shift back to its initial value ( 47.81 k ) within the experimental accuracy of @xmath11 0.02 k. these data demonstrate that the pressure dependence of @xmath4 is clearly nonlinear . the fact that @xmath20 depends on the temperature at which the pressure is changed gives evidence for relaxation behavior in overdoped hg-1201 , as discussed in the introduction ; the time dependence of this relaxation will be examined below . from fig .
2a the initial slope near 1 bar for pressure changes at low temperature is @xmath41 -1.15 @xmath11 0.52 k / gpa . to establish whether the above nonlinearity in @xmath20 is an accidental occurrence or is characteristic for overdoped hg-1201 samples , a second overdoped sample ( b ) with @xmath6 40.0 k from a different synthesis
was studied . as seen in fig .
2b , pressure was applied successively at 298 k ( open circles ) and released in steps ( closed circles ) at low temperature ( 65 k ) .
this procedure was then repeated ( open and closed triangles ) to check the reproducibility of the data .
evidence for nonlinearity in @xmath37 is also seen in fig .
2b for this sample ; however , the pressure above which the change in slope occurs appears to be shifted to higher pressures for sample b ( @xmath27 0.5 gpa ) relative to sample a ( @xmath27 0.4 gpa ) . for sample b we find the initial slopes @xmath38 -0.66 @xmath11 0.18 k / gpa and @xmath41 -0.57 @xmath11 0.30
k / gpa which have approximately half the magnitude of those for the less strongly overdoped sample a. the magnitude and even the sign of @xmath42 differ from the results of cao et al .
@xcite on similarly overdoped hg-1201 ; the reason for this discrepancy is unknown , but may arise from differences in the pressure medium ( he versus flourinert ) , pressure technique , details of sample preparation or measurement technique ( ac susceptibility versus resistivity ) .
interestingly , the data of cao et al .
@xcite for strongly overdoped hg-1201 reveal a nonlinearity in @xmath20 near 0.5 gpa , but the sign of the change in slope @xmath21 is opposite to ours .
we would now like to determine the temperature at which the relaxation behavior in the overdoped hg-1201 sample a initiates .
the sample studied ( second piece from the same pellet ) exhibited @xmath6 47.62 k at ambient pressure ( pt . 1 in fig .
3 ) . applying and holding 0.67 gpa pressure at 298 k for 1 h , @xmath4 decreased to pt . 2 ( 47.18 k ) .
immediately following the @xmath4 measurement , the pressure was fully released at 50 k , yielding @xmath6 47.32 k ( pt . 3 in fig .
3 ) , which lies 0.3 k lower than the initial value at ambient pressure .
the value of @xmath4 is found not to change following anneals for 1 h at 100 k ( pt .
4 ) and then for another hour at 200 k ( pt . 5 ) .
however , @xmath4 is seen to relax back toward its initial value following 1 h ( pt . 6 ) and then 3 h ( pt .
7 ) anneals at 298 k. the data in fig .
3 thus show that relaxation effects are only activated for temperatures above 200 k. we can not exclude the possibility that relaxation phenomena occur at temperatures below 50 k since in the present studies the measurement of @xmath4 itself constitutes annealing the sample at @xmath27 50 k. we now examine the time - dependence of the relaxation phenomena for overdoped hg-1201 at three different pressures . in these measurements a third piece of sample a was studied with @xmath6 47.81 k at ambient pressure ; full shielding was exhibited @xcite . a pressure of 0.61 gpa was applied and held 1 h at 298 k before cooling to measure @xmath43 47.38 k. after the full release of pressure at 48 k , we find @xmath6 47.58 k ( pt . 1 in fig .
4a ) . the sample
was then heated to 265 k and held for 1 h , before cooling rapidly to measure @xmath6 47.64 k ( pt . 2 in fig .
subsequent annealing at 265 k for longer times gives the @xmath4 values represented by points 3 to 7 in fig . 4a .
the fit to the data , which yields a relaxation time @xmath44 3.4 h , is discussed below .
the difference of 70 mk between the value of @xmath4 at the outset of the experiment ( 47.81 k ) and the final value ( 47.74 k ) is likely due to the different annealing temperature ( here @xmath45 265 k versus 298 k for the initial measurement ) .
we previously observed a similar effect in yba@xmath18cu@xmath1o@xmath10 @xcite .
as we discuss in more detail below , a lower annealing temperature leads to a higher state of oxygen order which increases the hole - carrier concentration in the cuo@xmath0 planes , shifting @xmath4 up for an underdoped sample or down for an overdoped sample , as observed .
our previous studies on yba@xmath0cu@xmath1o@xmath10 @xcite revealed that the relaxation behavior itself is strongly pressure dependent . to address this question for the present overdoped hg-1201 sample a
, a pressure of 0.62 gpa was applied and held at 298 k for 1 h. following the partial release of pressure to 0.13 gpa at 53 k , @xmath4 was found to be 47.41 k ( pt . 1 in fig .
4b ) . subsequent annealing at @xmath46 265 k ( p @xmath30 0.15 gpa at this temperature ) for various time periods up to a total of 145 h gives the relaxation behavior represented by points 2 to 8 in fig .
4b . a cursory comparison of the data in figs .
4a and 4b reveals that the relaxation time @xmath47265 k ) has become longer under pressure . to determine @xmath48 at higher pressures , the sample
was then subjected to 0.7 gpa pressure at 298 k for 1.6 h which was then partially released to 0.25 gpa at 52 k ( pt . 1 in fig .
because of the rapid increase in @xmath48 with pressure , the annealing temperature was increased to 285 k to speed up the relaxation .
the sample was annealed up to a total time of 42.5 h ( pts . 2 to 6 in fig .
we now analyze the time - dependent relaxation behavior of @xmath4 in fig . 4 in terms of the stretched exponential @xcite @xmath49\exp \left\ { -\left ( \frac{t}{\tau } \right ) ^{^{\alpha } } \right\ } , \ ] ] where @xmath50 is the transition temperature at a given pressure after infinite time , @xmath51 is the initial value of the transition temperature at a given pressure before relaxation begins , @xmath48 is thetemperature- and pressure - dependent relaxation time , and @xmath52 is the stretched exponent . in a previous analysis of relaxation behavior on a strongly underdoped y-123 single crystal @xcite , where the pressure - induced shifts in @xmath4 were large and the data could be analyzed to high accuracy , we found @xmath53 = 0.6 , independent of the pressure . for the present overdoped hg-1201 compound
the measured shifts in @xmath4 are much smaller , resulting in relatively large error bars . in the following we use @xmath53 = 0.6 @xcite and set the transition temperature at zero time , @xmath51 , to the value measured immediately after the pressure release at low temperatures , before any relaxation could take place .
a fit with eq .
( 1 ) to the ambient - pressure relaxation data at 265 k in fig .
4a ( solid line ) is seen to describe the data well within experimental error , yielding the parameters @xmath54 k and @xmath47265
k@xmath55 h. a second experiment ( not shown ) under similar conditions gives @xmath47265 k@xmath56 h , in agreement with the previous value .
a fit to the relaxation data at 0.15 gpa and 265 k in fig .
4b yields @xmath57 k and @xmath47265 k@xmath58 h. this relaxation time is distinctively longer than the ambient pressure values , implying that the relaxation slows down rapidly under pressure . since the relaxation at 0.30 gpa and 285 k in fig .
4c proceeded somewhat more rapidly than anticipated , a small correction to the annealing time is necessary to account for the finite time spent warming(cooling ) the sample to(from ) the annealing temperature .
based on the relaxation time obtained from a preliminary fit to the data with eq .
( 1 ) , we estimate that this correction amounts to only @xmath27 4 min . the corrected data in fig .
4c are now fit with eq .
( 1 ) , yielding @xmath59 k and a relaxation time @xmath47285 k@xmath60
h. we now study the pressure dependence of @xmath4 on an underdoped hg-1201 sample with @xmath4(1 bar ) @xmath8 47.55 k , an almost identical value to that for the overdoped sample a. since the superconducting onset is less sharp than that for the overdoped sample , changes in @xmath4 under pressure can not be determined as accurately ( @xmath11 50 mk versus @xmath11 20 mk ) . following the initial measurement of @xmath4 at ambient pressure ( pt . 1 in fig .
5 ) , a pressure of 0.79 gpa was applied at 298 k and held for 15 minutes before cooling down to measure @xmath6 48.55 k ( pt . 2 in fig . 5 ) ; the pressure at this temperature was 0.71 gpa , yielding the pressure derivative @xmath38 + 1.41 @xmath11 0.14
k / gpa , a value somewhat less than that ( + 2 k / gpa ) reported for underdoped hg-1201 by cao et al .
pressure was then successively decreased at temperatures below 90 k and @xmath4 immediately remeasured each time ( pts . 3 to 9 in fig .
5 ) , revealing a linear dependence of @xmath4 on pressure @xmath61 + 1.15 @xmath11 0.05 k / gpa . after the pressure was released ( pt .
9 ) , @xmath4 was found to lie 150 mk higher than the initial value ( pt .
1 ) . annealing the sample at 298 k for a total of 4 h ( pts . 10 - 12 ) fully restored the initial @xmath4 value .
annealing this sample for extended time periods at rt apparently causes some sample degradation , as evidenced by irreversible changes in the shape of the superconducting transition .
a strong degradation of underdoped hg-1201 samples stored in air was recently reported by peacock et al .
@xcite in which @xmath4 increased from 30 k to 75 k over the time span of 6.5 weeks following their synthesis .
this gradual shape change in the transition and the smallness of the relaxation effect itself precluded a detailed measurement of the time , temperature and pressure dependence of the relaxation phenomena for this underdoped sample ; from the data , however , we can give a rough estimate of the relaxation time at ambient pressure , namely @xmath47298 k ) @xmath33 1 h.
pressure - induced relaxation effects have been observed in several high-@xmath62 superconductors , including y-123 , tl-2201 and tl-1212 @xcite .
very recent studies on underdoped y-123 have shown that @xmath20 to 17 gpa depends markedly on whether the pressure is applied at ambient or low temperatures @xcite . in oxide superconductors , therefore , the measured dependence of @xmath4 on pressure may depend in a sensitive manner on the detailed pressure / thermal history of the sample .
the high pressure environment does not _
create _ the relaxation processes , but rather makes those already present under ambient conditions more readily visible than in conventional temperature - quench experiments .
theoretical calculations of the properties of matter , including their superconductivity , are generally carried out for a given fixed set of lattice parameters and atomic coordinates ; pressure effects are generally estimated under the assumption that the change in properties , like @xmath63 is a function of the anisotropic decrease in the lattice parameters , with all atomic separations scaling accordingly . before a measured @xmath20-dependence
can be compared in a quantitative way with theory , therefore , it should be purged of all influences from relaxation behavior .
since most high - pressure cells only allow the pressure to be changed at room temperature , many results from the large body of high - pressure studies on high-@xmath4 materials do contain relaxation effects and thus must be repeated using high - pressure technology similar to ours to separate intrinsic from relaxation effects . in the oxide superconductors the measured value of the transition temperature @xmath64 is a function of both the doping level @xmath22 and the sample s detailed pressure / temperature history ( i.e. the length of time @xmath65
the sample is annealed at a given temperature @xmath66 following a change in pressure @xmath67 ) ; it can be represented by the expression @xmath68 the first term on the right side gives the ideal or `` intrinsic '' change in @xmath17 which occurs immediately following a change in pressure at any temperature , whereas the second `` relaxation '' term takes into account the change in @xmath4 from pressure- , temperature- and time - dependent relaxation effects .
the simplest way to determine @xmath69 the dependence which can be compared most directly with theory , is to carry out the entire experiment at a temperature low enough to eliminate ( freeze out ) all relaxation processes , so that @xmath70 0 . as seen in fig .
3 , for hg-1201 relaxation processes appear to be frozen out for @xmath71 200 k. the two distinct `` intrinsic '' and `` relaxation '' contributions to @xmath72 for underdoped hg-1201 can be easily recognized in fig . 5 where the application of 0.79 gpa pressure at rt is seen to cause a larger change in @xmath4 than the pressure release below 90 k , corresponding to the pressure derivatives @xmath73 k / gpa and @xmath74 k / gpa , respectively .
whereas in the _ underdoped _ sample the change in @xmath4 with pressure is linear , in figs .
2a and 2b it is seen that for _ overdoped _ hg-1201 there is a distinct change in the slope of @xmath20 near 0.4 gpa for sample a and near 0.5 gpa for sample b. this evident nonlinearity is retained in the `` intrinsic '' @xmath75 dependence obtained by releasing the pressure at low temperatures .
this gives evidence that this nonlinearity is an intrinsic effect and does not have its origin in relaxation effects which occur at temperatures above 55 k , but may be due to a structural change . as pointed out previously ,
cao et al .
@xcite also observe a nonlinearity in @xmath20 near 0.5 gpa in strongly overdoped hg-1201 , but the change in slope is opposite in sign to the present results .
chen et al .
@xcite report a negative slope change in @xmath20 near 0.5 gpa in nearly optimally doped hg-1212 .
recent neutron powder diffraction experiments under high pressure on hg-1201 at variable oxygen content find no evidence for a structural phase transition to 1 gpa @xcite ; this study , however , did not include heavily overdoped samples with @xmath76 80 k , as in both the present experiments and those of cao et al .
further high - pressure structural studies are required to clarify the origin of the nonlinearity in @xmath20 .
the nonlinear @xmath20 dependence and the notable change in @xmath21 in the present experiments on the two heavily overdoped hg-1201 samples are but two of a number of indications that sufficiently overdoped hg-1201 ( where @xmath77 80 k ) possesses basic ( structural ) differences to underdoped hg-1201 .
cao et al .
@xcite have shown that the initial pressure derivative @xmath21 changes radically in the overdoped regime for @xmath78 80 k. xiong et al .
@xcite note that the @xmath79 lattice constant increases monotonically with oxygen content over the entire underdoped and partially overdoped regime , until @xmath16 80 k is reached , remaining constant thereafter ; the valence - bond sum technique for estimating carrier concentration also gives anomalous results for an oxygen content above this critical value .
xiong et al .
@xcite suggest that there may be two different types of lattice sites available for oxygen defects .
in fact , recent powder neutron diffraction experiments on hg-1201 by jorgensen et al .
@xcite over a wide range of oxygen defect concentrations suggest that there are two competing oxygen defects in the hg layer , an o3 defect in the center of the unit cell and an o4 defect nearer the cell s edge .
the relative concentrations of these defects switches with increasing @xmath22 upon passing from the underdoped region and through the maximum @xmath4 , where o4 is the dominant defect , to the overdoped region where , at @xmath6 80 k , o3 suddenly becomes the dominant defect .
this `` change of the defect '' is mirrored in the dependence of @xmath4 on @xmath22 which deviates from the canonical inverted parabolic behavior .
of particular relevance to the present experiments is the fact that a sudden _ increase _ in the unit cell volume by @xmath27 1% accompanies this `` change of the defect '' o4 @xmath80o3 at @xmath6 80 k @xcite .
since o3 should be the sole defect in our strongly overdoped samples , this implies that if sufficiently high pressure is applied , the o4 defect should become energetically favored over o3 . in this picture
the nonlinearity in @xmath37 would signal a sudden changeover from a dominant o3 defect to o4 .
the fact that the critical onset pressure for sample b is somewhat larger ( 0.5 gpa ) than for sample a ( 0.4 gpa ) supports this picture since sample b is more strongly overdoped and thus the o3 defect more stable .
neutron diffraction experiments under high pressure conditions would be desirable to clarify the anomalous but interesting overdoped state in hg-1201 .
it is interesting to note that for under- and overdoped hg-1201 with @xmath81 49 k the initial `` intrinsic '' slopes @xmath82 -1.15 @xmath11 0.05 k / gpa and -1.15 @xmath11 0.52 k / gpa , respectively , are equal in magnitude but opposite in sign . at first glance
this result would appear to support a simple charge - transfer model @xcite in which @xmath83 and @xmath84 constant 0 . however
, this would lead to the expectation that @xmath85 0 for an optimally doped sample ( where @xmath86 0 ) , contrary to the experimental result that @xmath87 + 1.75 @xmath11 0.1 k / gpa @xcite .
we note that the intrinsic pressure dependence for the more strongly overdoped sample b is only half the magnitude of that for sample a ; the charge transfer model would lead to the expectation that @xmath88 should be _
greater _ in magnitude , contrary to experiment .
it is certainly possible that pressure - dependent shifts in the defect structure in hg-1201 play a role here
. further measurements of @xmath37 together with systematic structural studies under high pressure are needed to resolve this issue . in order to better characterize the relaxation behavior at different pressures and temperatures in overdoped hg-1201
, we calculate the pressure - dependent activation energy @xmath89 from the measured values of the relaxation time @xmath90 at a given temperature using the arrhenius law we set the attempt period @xmath92 equal to the value ( @xmath93 1.4@xmath2810@xmath94 s ) found in earlier temperature - quench experiments on y-123 @xcite . using the values of the relaxation times at ambient pressure from the previous section ,
the corresponding activation energy for overdoped ( sample a ) and underdoped hg-1201 can be calculated using eq .
3 , yielding @xmath951 bar ) @xmath8 0.84 ev and 0.90 ev , respectively .
these values are quite comparable to each other and to activation energies determined from relaxation studies on y-123 ( 0.97 ev ) @xcite , tl-2201 ( 0.72 ev ) @xcite and tl-1212 ( 0.86 ev ) @xcite , as well as to oxygen tracer diffusion studies on y-123 and other oxide superconductors @xcite
this fact , and the known sensitivity of the relaxation behavior to the oxygen content @xcite , make it likely that the above relaxation effects originate in the oxygen sublattice . to our knowledge
, no attempt has yet been made to compute the energy barrier for oxygen diffusion in hg-1201 ; in a recent model calculation islam and winch @xcite have estimated @xmath951 bar ) @xmath33 0.68 ev for the migration of interstitial o(4 ) oxygen in the three - layer compound hg-1223 , a value roughly comparable to our present result for hg-1201 .
it is interesting to note that no relaxation whatsoever has been observed in @xmath4 for optimally doped hg-1201 @xcite or , for that matter , for any other optimally doped oxide superconductor studied to date @xcite .
does this mean that in optimally doped samples relaxation phenomena do not occur ? certainly not !
the absence of relaxation phenomena in @xmath4 for optimally doped samples can be readily understood if the change in @xmath4 due to relaxation phenomena originates primarily from a change in the carrier concentration @xmath96 in the cuo@xmath0 plane(s ) ; hall effect studies on tl-2201 support this conclusion @xcite . in an optimally doped sample the value of @xmath4 is maximized as a function of @xmath97 i.e. @xmath98 is at an extremum where @xmath99 0 , so that @xmath100 0 . in optimally doped samples ,
therefore , relaxation processes in the oxygen sublattice initially cause no change in @xmath63 but should have a measurable effect on other properties such as the electrical resistivity , hall coefficient , lattice parameters , etc . in this regard
further studies on the hg - oxides would be of particular interest since optimal doping occurs at a _ finite _ concentration of oxygen defects , in contrast to y-123 and tl-1201 which are nearly devoid of oxygen defects when @xmath4 is at its maximum value .
an important question remaining is to identify the mechanism by which the relaxation processes change the carrier concentration in hg-1201 .
for y-123 veal and paulikas @xcite proposed that the local rearrangement of oxygen anions in the cuo chain layer causes the valence of the ambivalent cu@xmath101 cations in this layer to change .
the maintainance of overall charge neutrality then dictates that the charge carrier concentration in the cuo@xmath0 layers must adjust accordingly . in tl-2201
the oxygen defects are incorporated interstitially in the tl@xmath0o@xmath18 double - layer @xcite ; some form of local ordering of these oxygen would then prompt some of the ambivalent tl@xmath102-cations in this layer to decrease their valence by drawing electron charge away from the cuo@xmath18 plane ( for thallium + 1 and + 3 are the preferred valence states ) .
the presence of ambivalent cations would not appear to be mandatory for charge transfer to occur , but probably serves to enhance it ; as pointed out in the introduction , relaxation phenomena have been reported in superoxygenated la@xmath18cuo@xmath24 @xcite , although this compound contains no ambivalent cations outside the cuo@xmath0 layer .
hg-1201 would appear to be a further such example , since xps studies @xcite on an optimally doped sample indicate that the valence of hg is 2 + .
further evidence against hg cation ambivalence in hg - oxides is given by chemical analysis on a series of hg-1212 samples near optimal doping which show that the average cu valence increases with oxygen content , but that the valence of hg is fixed at 2 + @xcite .
two possible mechanisms for the observed weak relaxation phenomena in hg-1201 are : ( 1 ) the degree of oxygen ordering may directly influence how the available hole carriers are distributed between the cuo@xmath18 planes and other structural elements @xcite ; ( 2 ) ambivalent cu@xmath103 cations on hg sites may switch their valence as oxygen defects rearrange .
indeed , studies by wagner et al .
@xcite show that there is a substitutional defect in hg-1201 in the form of a partial substitution of cu ions on the hg site ; the same overdoped hg-1201 sample a used in the present experiments was found to contain a substitution of @xmath104 8% of cu on the hg site @xcite which could conceivably lead to modest relaxation effects of the order @xmath27 8% of that expected for full substitution .
it is interesting to note that in the present experiment on overdoped hg-1201 ( sample a ) the magnitude of the observed relaxation derivative ( @xmath105 -0.3 k / gpa ) is only @xmath106 10% of that for y-123 @xcite with a similar value of @xmath17 .
a critical test would be to study non optimally doped hg-1201 samples devoid of cu substitution ; if the second scenario is correct , there would then be no relaxation in @xmath4 . using the values of @xmath90 for overdoped hg-1201 ( sample a ) from the previous section
, @xmath89 is calculated using eq .
3 and displayed in fig .
we ignore any change in @xmath92 with pressure since its anticipated decrease under pressure is very small @xcite . in this figure
the value of the pressure near the annealing temperature @xmath107 rather than the value near @xmath63 is used since below 250 k only insignificant relaxation occurs .
assuming that , as found previously for y-123 @xcite , @xmath7 is a linear function of pressure , we obtain @xmath108 ( + 114 @xmath11 50 ) mev / gpa ( dashed line in fig . 6 ) . to our knowledge ,
the only other such study on a superconducting oxide was on strongly underdoped yba@xmath0cu@xmath1o@xmath10
@xcite where we found @xmath109 ( + 44 @xmath11 2 ) mev / gpa , approximately half the present value for hg-1201 . the pressure dependence of relaxation behavior is normally analyzed in terms of an activation volume @xmath110 which can be calculated from the pressure dependence of the activation energy @xcite @xmath111 where @xmath112 is the change in volume due to the formation of a vacancy or interstitial
, @xmath113 is the volume change due to the migration of the diffusing atom , and @xmath114 is avogadro s number . if we assume that the oxygen defects involved in the relaxation phenomena are already present and need not be created , then @xmath112 = 0 .
as discussed above , two distinct types of oxygen defect have been found in the hgo layer in hg-1201 @xcite , with a transition from one type to the other near the optimal doping concentration .
the nonlinearity in @xmath20 seen in fig .
2 near 0.4 gpa for the overdoped samples may signal a transition from the o3 to o4 defect type . since in the present experiments the time dependence of the relaxation behavior
was only studied for pressures below 0.3 gpa , we assume that only a single oxygen defect type is responsible for the relaxation , so that @xmath115 . from eq .
( 4 ) we now estimate from the above value of @xmath116 the molar activation volume to be @xmath110 @xmath117 + ( 11.0 @xmath11 4.8 ) cm@xmath29/mol which is approximately twice the molar volume of the diffusing o@xmath119 ion ( 5.78 cm@xmath29/mol @xcite ) . in our recent study on underdoped yba@xmath0cu@xmath1o@xmath10 an activation volume of
+ 4.2 cm@xmath29/mol @xcite was obtained .
a positive migration volume @xmath113 implies that the lattice increases in volume as a diffusing ion passes over a saddle point in moving from one site to another .
it is physically reasonable that diffusion slows under pressure ( @xmath120 0 ) because squeezing the lattice together leaves less space through which the interstitial atoms can diffuse . indeed , in ionic compounds @xcite , such as the oxide superconductors , a positive activation volume is normally observed .
however , an enhancement of the diffusion rate under pressure ( negative activation volume ) may be observed in covalently bonded substances , such as amorphous si @xcite .
recently , the activation volume for both hg-1201 and y-123 has been calculated @xcite using a hard sphere model to estimate the local expansion in the lattice as the oxygen defect travels along its diffusion path . for both systems the measured values of @xmath110
are underestimated by a factor of two to three in this model . in summary , the present experiments on hg-1201 give evidence for weak relaxation phenomena in the transition temperature for both under- and overdoped samples ; these relaxation effects are opposite in sign but roughly equal in magnitude and involve similar activation energies .
this result , plus the absence of relaxation effects in @xmath4 for optimally doped hg-1201 @xcite , support the view that relaxation phenomena influence @xmath4 via the transfer of charge in or out of the cuo@xmath0 planes .
the relaxation time at ambient pressure is found to be approximately 4 h at 265 k and increases strongly with pressure .
this pressure dependence can be expressed in terms of an activation volume equal to + 11.0 cm@xmath29/mol .
an anomalous nonlinearity in @xmath20 is observed for the overdoped samples near 0.4 gpa which may indicate a transition from one oxygen defect type to another .
the work at washington university ( ss and jss ) is supported by the national science foundation under grant dmr 98 - 03820 .
part of this work supported by the nsf office of science and technology centers , grant no .
dmr 91 - 20000 ( oc and bd ) and us department of energy , basic energy sciences - material sciences , contract no .
w-31 - 109-eng-38 ( jdj and dgh ) .
useful discussions with a .- k .
klehe are acknowledged .
j. s. schilling , j. diederichs , s. klotz and r. sieburger in : _ magnetic susceptibility of superconductors and other spin systems _ , eds .
r. a. hein , t. l. francavilla and d. h. liebenberg ( plenum , new york , 1991 ) p. 107 . a detailed analysis of the optimal value of @xmath122 as carried out for y-123 ( see ref .
31 ) is not possible here because of the small shifts in @xmath4 .
nevertheless , @xmath53 = 0.6 consistently gives the best fits for hg-1201 . in a simple grneisen model , @xmath123 -2%/gpa , where @xmath124 1%/gpa is the compressibility and
@xmath125 2 is the grneisen constant .
this change in @xmath92 is negligible compared to the measured change in @xmath126 itself , where @xmath127 + 500%/gpa . *
fig . 1 . * real part of the ac - susceptibility versus temperature for overdoped hg-1201 ( sample a ) at ambient pressure .
data points become very dense above 40 k. solid line is guide to the eye .
the inset shows a blow - up of the data near the superconducting onset temperature @xmath4 . *
* dependence of @xmath4 on hydrostatic pressure for two samples of overdoped hg-1201 : ( a ) @xmath1281 bar ) @xmath8 47.8 k
( sample a ) ; pressure was successively increased at 298 k ( o ) and then successively decreased at either 298 k ( @xmath129 ) or at temperatures 55 k ( @xmath130 ) .
( b ) @xmath1281 bar ) @xmath8 40.0 k ( sample b ) ; pressure was successively increased at 298 k ( o ) and then successively decreased at temperatures 65 k ( @xmath131 ) , with numbers giving the order of measurement .
pressure was then reapplied ( @xmath132 ) at 298 k and released ( @xmath133 ) at t 65 k , as before .
see text for full details .
solid lines are guides to the eye .
* dependence of @xmath4 on annealing temperature t@xmath134 in overdoped hg-1201 ( sample a ) following the application of pressure at 298 k and its removal at 50 k. relaxation effects are only observed for temperatures above 200 k. numbers give order of measurement . dashed and solid lines are guides to the eye . *
* relaxation data on overdoped hg-1201 ( sample a ) at three different pressures .
the solid lines give fits using eq .
( 1 ) with given values of @xmath48 .
numbers give order of measurement .
see text for details .
* dependence of @xmath4 on pressure in underdoped hg-1201 where 0.7 gpa pressure is applied at 298 k but released at temperatures below 90 k. numbers give order of measurement . dashed and solid lines are guides to the eye .
see text for details . | in previous studies on a number of under- and overdoped high temperature superconductors , including yba@xmath0cu@xmath1o@xmath2 and tl@xmath0ba@xmath0cuo@xmath3 , the transition temperature @xmath4 has been found to change with time in a manner which depends on the sample s detailed temperature and pressure history .
this relaxation behavior in @xmath4 is believed to originate from rearrangements within the oxygen sublattice . in the present high - pressure studies on hgba@xmath0cuo@xmath5 to 0.8 gpa we find clear evidence for weak relaxation effects in strongly under- and overdoped samples ( @xmath6 40 - 50 k ) with an activation energy @xmath7(1 bar ) @xmath8 0.8 - 0.9 ev . for overdoped hgba@xmath0cuo@xmath5 @xmath9 increases under pressure more rapidly than previously observed for yba@xmath0cu@xmath1o@xmath10 , yielding an activation volume of + 11 @xmath11 5 @xmath12 ; the dependence of @xmath4 on pressure is markedly nonlinear , an anomalous result for high-@xmath4 superconductors in the present pressure range , giving evidence for a change in the electronic and/or structural properties near 0.4 gpa .
[ theorem]acknowledgement [ theorem]algorithm [ theorem]axiom [ theorem]claim [ theorem]conclusion [ theorem]condition [ theorem]conjecture [ theorem]corollary [ theorem]criterion [ theorem]definition [ theorem]example [ theorem]exercise [ theorem]lemma [ theorem]notation [ theorem]problem [ theorem]proposition [ theorem]remark [ theorem]solution [ theorem]summary |
in the cb model , long - duration grbs and their ags are produced in core collapse supernovae by jets of highly relativistic `` cannonballs '' that pierce through the supernova shell . the ag the persistent radiation in the direction of an observed grb has three origins : the ejected cbs , the concomitant sn explosion , and the host galaxy .
these components are usually unresolved in the measured `` grb afterglows '' , so that the corresponding light curves and spectra refer to the cumulative energy flux density : @xmath6 the contribution of the candidate host galaxy depends on the angular aperture of the observations and it is usually determined at late times , when the cb and sn contributions have become negligible .
core - collapse supernovae ( sn of types ii / ib / ic ) are far from being standard candles .
but if their explosions are fairly asymmetric as they would be if a fair fraction of them emit two opposite jets of cbs much of the variability could be a reflection of the varying angles from which we see their non - spherically expanding shells . exploiting this possibility to its extreme , we use sn1998bw as an ansatz standard candle ( dado et al .
2001a and references therein ) .
let the energy flux density of sn1998bw at redshift @xmath1 ( galama et al . 1998 ) be @xmath7 $ ] .
for a similar sn placed at a redshift @xmath8 : @xmath9 = } & & { \rm{1+z \over 1+z_{bw}}\ ; { d_l^2(z_{bw})\over d_l^2(z)}}\ , \times\nonumber \\ & & { \rm f_{bw}\left[\nu\,{1+z \over 1+z_{bw}},t\ , { 1+z_{bw } \over 1+z}\right]\ ; a(\nu , z)}\ , , \label{bw}\end{aligned}\ ] ] where @xmath10 is the luminosity distance km/(s mpc ) , @xmath11 and @xmath12 .
] and @xmath13 is the extinction along the line of sight .
the selective extinction in the galaxy in the direction of grb 010921 is @xmath14 ( schlegel et al .
one can use standard conversion of e(b - v ) to the attenuation in our galaxy , due to dust along the line of sight to the grb at different wave lengths .
the selective extinction in our galaxy in the direction of grb 010921 is expected to increase the spectral slope of the observed optical light by @xmath15 .
the unextinct late time contribution of a cb to the grb afterglow is given by : @xmath16^{3\alpha-1}\;[\delta(t)]^{3+\alpha}\ , \nu^{-\alpha } , \label{fluxdensity2}\ ] ] where @xmath17 is a normalization constant ( see dado et al .
2001a for its theoretical estimate ) , @xmath18 is the spectral index of the synchrotron radiation from fermi - accelerated electrons in the cb whose acceleration rate is in equilibrium with their synchrotron cooling rate , @xmath19 is the lorentz factor of the cb , and @xmath20 is its doppler factor : @xmath21 whose approximate expression is valid for small observing angles @xmath22 and @xmath23 , the domain of interest for grbs . for an interstellar medium of constant baryon density @xmath24 ,
the lorentz factor , @xmath25 , as a function of observer s time , is given by ( dado et al .
2001a ) : @xmath26\nonumber\\ \rm c&\equiv&\rm
\left[{2\over b^2 + 2\,\theta^6+b\,\sqrt{b^2 + 4\,\theta^6}}\right]^{1/3 } \nonumber\\ \rm b&\equiv&\rm { 1\over \gamma_0 ^ 3}+{3\,\theta^2\over\gamma_0}+ { 6\,c\ , t\over ( 1+z)\ , x_\infty } \label{cubic}\end{aligned}\ ] ] where @xmath27 , and @xmath28 characterizes the cb s slow - down in terms of @xmath29 : its baryon number , and @xmath30 : its radius ( it takes a distance @xmath31 for the cb to half its original lorentz factor ) . after a correction for selective extinction in our galaxy ( @xmath14 magnitudes in the direction of grb 010921 ) the photometric measurements of its optical ag yield a spectral slope @xmath32 ( park et al .
2001 ; price et al .
this slope is much steeper than the normally observed spectral slopes of grb afterglows : @xmath33 at early time ( the first couple of days ) and @xmath34 later .
we interpret this steepening as the effect of selective extinction in the host galaxy ; a similarlly strong steepening of the spectral index , also consistent with strong extinction in the host galaxy , has been observed in the ag of a few grbs , e.g. , grb 990712 ( bjornsson et al . 2001 ) and grb 000926 ( harrison et al .
supportive evidence for strong extinction in the host galaxy of grb 010921 is provided by the spectral index of its continuum light and by the relative intensity of its measured @xmath35 to @xmath36 emission lines ( price et al .
2002a ) , which is different from the expected intrinsic ratio of 2.87 ( mathis 1983 ) and yields an optical depth at @xmath36 due to dust of 1.51 .
assuming an intrinsic @xmath37 for the late - time ag ( in agreement with all observed ags of grbs of known redshift , see dado et al .
2001a ) , the best - fitted parameters to the ag of grb 011121 are : @xmath38 , @xmath39 , and @xmath40 . in figs . 1 to 3
, we show the fitted light curves and the observations in the v , r and i bands , assuming a sn1998bw - like contribution at @xmath4 .
the relative intensities were adjusted to fit the early time data .
the difference between the expected intensity ratios for a spectral slope -1.1 and the observed relative intensities were used to correct the sn1998bw contribution for selective extinction in the host galaxy and in the milky way .
the selective extinction in our galaxy and in the host was also used to estimate roughly the overall attenuation normalization : @xmath41 magnitudes in the r band .
the fits are clearly very satisfactory .
the spectral index of the optical ag for most grbs after a couple of days is approximately @xmath42 ( e.g. , simon et al . , 2001 ) . when dominated by a supernova contribution , the optical afterglows of grbs become much more red ( see , e.g. , reichart 1999 ; bloom et al . 1999 ) .
the spectral index of the optical ag of grb 010921 steepened to @xmath43 on day 35 and to @xmath44 on day 46 ( price et al .
2002b ) as expected when the ag is dominated by a supernova like sn1998bw .
this is demonstrated in figs .
4 and 5 where we compare the colours of sn1998bw at @xmath4 after extinction in the host galaxy and in ours and the colours of the ag of grb 010921 , as measured with hst on days 35 and 46 after burst .
the colours observed by hst are in good agreement with the expectation .
contrary to the questionable and inconclusive fireball - model analysis of the optical afterglow of grb 010921 , our analysis , based on the cannonball model , shows strong evidence for the contribution of a supernova akin to sn1998bw , seen at the grb s redshift . the evidence is clearly there in all of the relevant observations
: the i , r and v light curves , and the spectra at the times when the supernova s contribution is dominant .
bjornsson , g. , et al . , 2001 , apj , 552 , 121l bloom , j.s . , et al . , 1999 ,
nature 401 , 452 bloom , j.s .
et al . , 2002 ,
gcn 1274 castro - tirado a.j .
& gorosabel , j. , 1999 , a&as 138(3 ) , 449 castro - tirado , a.j . , et al . , 2001
, a&a 370 , 398 dado , s. , dar , a. & de rjula , a. , 2001a , astro - ph/0107367 + , accepted for publication in a&a dado , s. , dar , a. & de rjula , a. , 2001b , astro - ph/0111468 dar , a. 1999a , a&as 138(3 ) , 505 dar , a. , 1999b , gcn circ .
346 dar , a. , 1998a , apj 500 , l93 dar , a. & de rjula , a. , 2000b , astro - ph/0012227 , + submitted to a&a dar , a. & plaga , r. , 1999 , a&a 349 , 259 djorgovski , s.g .
, et al . , 2001 , gcn circ .
1108 fox , d.w .
, et al . , 2001 , gcn circ .
1099 galama , t.j . , et al . , 1998 ,
nature 395 , 670 garnavitch , p.m. et al .
, gcn circ .
1273 galama , t.j . , et al . , 2000 ,
apj 536 185 ghisellini , g. astro - ph/0111584 harrison , f.a .
et al . , 2001 ,
apj 559 , 123 henden , a. , et al .
2001 , gcn circ .
1100 hjorth , j. , et al .
, 2000 , apj 534 , 147l holland , s. , et al . , 2001 , a&a 371 , 52 hurley , k. , et al . , 2001 ,
1097 klose , s. stecklum b. , 2001 , gcn circ .
1113 kulkarni , s. et al .
, 2002 , gcn 1276 lazzati , d. et al . , 2001 , a&a 378 , 996 mathis , j.s .
1983 , apj 267 , 119 park , h.s . , et al . , 2001 , astro - ph/0112397 price , p.a . , et al
. , 2001 , gcn circ .
1107 price , p.a .
, et al . , 2002a , astro - ph/0201399 price , p.a . , et al . , 2002b , gcn circ 1259 reichart , d.e . , 1999 ,
apj 521 , l111 ricker , g. , et al . , 2001 ,
1096 sahu , k.c . , et al . , 2000 ,
apj 540 , 74 schlegel , d.j . ,
finkbeiner , d.p . &
davis , m. , 1998 , apj 500 , 525 shaviv , n.j . &
dar , a. , 1995 , apj 447 , 863 simon , v. , et al . , 2001 ,
astro - ph/0108416 sokolov , v.v .
, et al . , 2001a , astro - ph/0102492 | we use the very simple and successful cannonball model ( cb ) of gamma ray bursts ( grbs ) and their afterglows ( ags ) to analyze the observations of the strongly extinct optical ag of grb 010921 with ground - based telescopes at early times , and with the hst at later time .
we show that grb 010921 was indeed associated with a 1998bw - like supernova at the grb s redshift .
the identity of the progenitors of gamma ray bursts is still debated .
it has been suggested that grbs are produced by highly relativistic jets ( e.g. , shaviv and dar 1995 , dar 1998 ) mostly from supernovae like sn1998bw ( dar and plaga 1999 , dar 1999a , dar and de rjula 2000 and references therein ) . evidence for a sn1998bw - like contribution to a grb afterglow ( dar 1999a ; castro - tirado & gorosabel 1999 ) was first found by bloom et al .
( 1999a ) for grb 980326 , but the unknown redshift prevented a quantitative analysis .
the ag of grb 970228 ( located at redshift @xmath0 ) appears to be overtaken by a light curve akin to that of sn1998bw ( located at @xmath1 ) , when properly scaled by their differing redshifts ( dar 1999b ; reichart 1999 ; galama et al .
2000 ) .
evidence of similar associations was found for grb 990712 ( hjorth et al .
2000 ; sahu et al .
2000 ; bjornsson et al . 2001 ) , grb 980703 ( holland et al .
2000 ) , grb 000418 ( dar and de rjula 2000 ) , grb 991208 ( castro - tirado et al . 2001 ) , grb 970508 ( sokolov et al .
2001 ) and grb 000911 ( lazzati et al .
2001 ) .
dar and de rjula ( 2000 ) and dado et al .
( 2001a ) have shown that the optical ag of all relatively nearby grbs with known redshift ( all the ones with @xmath2 ) contain evidence or hints for a sn1998bw - like contribution to their optical ag , suggesting that most and perhaps all of the long duration grbs are associated with 1998bw - like supernovae , but more cases are needed before a firm conclusion can be reached .
on 2001 september 21.21934 ut a bright gamma ray burst ( grb 010921 ) was detected and localized ( ricker et al .
2002 ) by the high energy transient explorer ( hete2 ) and the interplanetary network ipn ( hurley et al .
2001 ) .
early observations of the error box of grb 010921 with large aperture telescopes found no evidence of an optical afterglow to the limit r @xmath3 20.5 ( fox et al .
2001 , henden et al .
2001 )
. however , follow - up observations by price et al .
( 2001 ) resulted in the discovery of its optical and radio ags .
this event was the first grb localized by hete2 which has resulted in the detection of an ag .
further observations by various telescopes ( park et al . 2002 , price 2002a and references therein ) have shown that the grb is located in a relatively nearby host galaxy with redshift @xmath4 ( djorgovski et al .
2001 ) and that its optical afterglow exhibits a fast temporal decline and has a large spectral slope , @xmath5 , after correcting for extinction in our galaxy , ( e(b - v)=0.148 ; schlegel et al .
1998 ) , in the direction of the grb .
due to its relative proximity and fast temporal decline , the ag of grb 010921 should have shown an sn1998bw - like contribution , after subtraction of its host galaxy contribution , if it was associated with a supernova akin to 1998bw , located at @xmath4 . in order to detect or constrain an underlying supernova in grb 010921 ,
observations with hst were obtained on 2001 october 26 , november 6 and 25 , and 2002 jan 4 by price et al .
( 2002b ) . from a fireball - model analysis of these observations , these authors identified a break in the optical light - curve around day 35 which they argued is due to the collimated ejecta , and concluded that the existence of a sn with a luminosity greater than 20% of sn 1998bw at @xmath4 is ruled out .
however , in this note , using the very successful cannonball model of grbs ( dar and de rjula 2000 ) and their afterglows ( dado et al .
2001a ) to analyze the light - curve and spectral evolution of the ag of grb 010921 , we show that there is strong and clear evidence that grb 010921 was associated with a supernova akin to 1998bw , at the grb s redshift . |
our main goal in this paper is to study global rates of convergence of the maximum likelihood estimator ( mle ) in one simple model for multivariate interval - censored data . in section 3
we will show that under some reasonable conditions the mle converges in a hellinger metric to the true distribution function on @xmath0 at a rate no worse than @xmath3 for @xmath4 for all @xmath5 .
thus the rate of convergence is only worse than the known rate of @xmath6 for the case @xmath7 by a factor involving a power of @xmath8 growing linearly with the dimension .
these new rate results rely heavily on recent bracketing entropy bounds for @xmath9dimensional distribution functions obtained by @xcite .
we begin in section [ sec : intcensonr ] with a review of interval censoring problems and known results in the case @xmath7 .
we introduce the multivariate interval censoring model of interest here in section [ sec : multintervalcensoring ] , and obtain a rate of convergence for this model for @xmath5 in theorem [ newmulthellingerratethm ] .
most of the proofs are given in section [ sec : proofs ] , with the exception being a key corollary of @xcite , the statement and proof of which are given in the appendix ( section [ sec : appendix ] ) .
finally , in section [ sec : othermodelsfurtherproblems ] we introduce several related models and further problems .
let @xmath11 on @xmath12 , and let @xmath13 on @xmath12 be independent of @xmath14 .
suppose that we observe @xmath15 i.i.d . as @xmath16 where @xmath17}$ ] .
here @xmath14 is often the time until some event of interest and @xmath18 is an observation time .
the goal is to estimate @xmath19 nonparametrically based on observation of the @xmath20 s . to calculate the likelihood
, we first calculate the distribution of @xmath21 for a general distribution function @xmath22 : note that the conditional distribution of @xmath23 conditional on @xmath18 is bernoulli : @xmath24 where @xmath25 . if @xmath26 has density @xmath27 with respect to some measure @xmath28 on @xmath12 , then @xmath16 has density @xmath29 with respect to the dominating measure ( counting measure on @xmath30 .
the nonparametric maximum likelihood estimator ( mle ) @xmath31 of @xmath19 in this interval censoring model was first obtained by @xcite .
it is simply described as follows : let @xmath32 denote the order statistics corresponding to @xmath33 and let @xmath34 denote the corresponding @xmath23 s
. then the part of the log - likelihood of @xmath35 depending on @xmath22 is given by @xmath36 where @xmath37 it turns out that the maximizer @xmath31 of ( [ loglikelihood ] ) subject to ( [ monconstr ] ) can be described as follows : let @xmath38 be the ( greatest ) convex minorant of the points @xmath39 : @xmath40 let @xmath41 denote the left - derivative of @xmath38 at @xmath42
. then @xmath43 is the unique vector maximizing ( [ loglikelihood ] ) subject to ( [ monconstr ] ) , and we therefore take the mle @xmath31 of @xmath22 to be @xmath44 with the conventions @xmath45 and @xmath46 . see @xcite or @xcite , pages 38 - 43 , for details .
@xcite initiated the study of @xmath31 and proved the following limiting distribution result at a fixed point @xmath47 .
( groeneboom , 1987 ) .
[ groeneboompointwiselimitdistributionrone ] consider the current status model on @xmath12 .
suppose that @xmath48 and suppose that @xmath22 and @xmath49 are differentiable at @xmath47 with strictly positive derivatives @xmath50 and @xmath51 respectively .
then @xmath52 where @xmath53 and @xmath54 where @xmath55 is a standard two - sided brownian motion starting from @xmath56 .
the distribution of @xmath57 has been studied in detail by @xcite and computed by @xcite .
@xcite show that the density @xmath58 of @xmath57 is log - concave .
@xcite ( see also @xcite ) obtained the following global rate result for @xmath59 . recall that the hellinger distance @xmath60 between two densities with respect to a dominating measure @xmath28 is given by @xmath61 ( van de geer , 1993 ) [ vdghellingerraterone ] @xmath62 .
now for any distribution functions @xmath22 and @xmath19 the ( squared ) hellinger distance @xmath63 for the current status model is given by @xmath64 and hence proposition [ vdghellingerraterone ] yields @xmath65 or @xmath66 .
for generalizations of these and other asymptotic results for the current status model to more complicated interval censoring schemes for real - valued random variables @xmath14 , see e.g. @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite .
our main focus in this paper , however , concerns one simple generalization of the interval censoring model for @xmath10 introduced above to interval censoring in @xmath0 .
we now turn to this generalization .
let @xmath67 on @xmath68 , and let @xmath69 on @xmath70 be independent of @xmath71 .
we assume that @xmath26 has density @xmath27 with respect to some dominating measure @xmath28 on @xmath0 .
suppose we observe @xmath72 i.i.d . as @xmath73 where @xmath74 is given by @xmath75}$ ] , @xmath76 .
equivalently , with a slight abuse of notation , @xmath77 where @xmath78 is a vector of length @xmath79 consisting of @xmath56 s and @xmath80 s and with at most one @xmath80 which indicates into which of the @xmath79 orthants of @xmath70 determined by @xmath81 the random vector @xmath71 belongs .
more explicitly , define @xmath82 .
then set @xmath83 for @xmath84 , so that @xmath85 and @xmath86 for @xmath87 .
much as for univariate current status data , @xmath71 represents a vector of times to events , @xmath81 is a vector of observation times , and the goal is nonparametric estimation of the joint distribution function @xmath19 of @xmath71 based on observation of the @xmath88 s .
see @xcite , @xcite , @xcite , and @xcite for examples of settings in which data of this type arises . to calculate the likelihood
, we first calculate the distribution of @xmath89 for a general distribution function @xmath22 : note that the conditional distribution of @xmath90 conditional on @xmath81 is multinomial : @xmath91 where @xmath92 and the probabilities @xmath93 , @xmath94 , @xmath95 are determined by the @xmath22 measures of the corresponding sets .
then our model @xmath96 for multivariate current status data is the collection of all densities with respect to the dominating measure @xmath97 given by @xmath98 for some distribution function @xmath22 on @xmath70 where @xmath95 and @xmath99 with @xmath100 . now
the part of the log - likelihood that depends on @xmath22 is given by @xmath101 and again the mle @xmath31 of the true distribution function @xmath19 is given by @xmath102 for example , when @xmath103 , we can write @xmath104 , @xmath105 , @xmath106 , and @xmath107 , and then @xmath108 thus @xmath109 note that @xmath110 where @xmath111\times [ 0,t_2 ] , \\ & & c_2 ( \underline{t } ) = [ 0,t_1]\times ( t_2 , \infty ) , \\ & & c_3 ( \underline{t } ) = ( t_1 , \infty ) \times [ 0,t_2 ] , \\ & & c_4 ( \underline{t } ) = ( t_1 , \infty ) \times ( t_2 , \infty ) .\end{aligned}\ ] ] characterizations and computation of the mle ( [ multivariateintcensmle ] ) , mostly for the case @xmath103 have been treated in @xcite , @xcite , and @xcite . consistency of the mle for more general interval censoring models has been established by @xcite .
for an interesting application see @xcite .
this example and other examples of multivariate interval censored data are treated in @xcite and and @xcite . for a comparison of the mle with alternative estimators in the case @xmath103 ,
see @xcite . an analogue of groeneboom s theorem [ groeneboompointwiselimitdistributionrone ]
has not been established in the multivariate case .
@xcite established an asymptotic minimax lower bound for pointwise convergence when @xmath103 : if @xmath19 and @xmath26 have positive continuous densities at @xmath112 , then no estimator has a local minimax rate for estimation of @xmath113 faster than @xmath6 . by making use of additional smoothness hypotheses
, @xcite has constructed estimators which achieve the pointwise @xmath6 rate , but it is not yet known if the mle achieves this .
our main goal here is to prove the following theorem concerning the global rate of convergence of the mle @xmath31 .
[ newmulthellingerratethm ] consider the multivariate current status model .
suppose that @xmath19 has @xmath114^d$ ] and that @xmath19 has density @xmath115 which satisfies @xmath116^d \label{conditionone}\end{aligned}\ ] ] where @xmath117 .
suppose that @xmath26 has density @xmath27 which satisfies @xmath118^d .
\label{conditiontwo}\end{aligned}\ ] ] then the mle @xmath119 of @xmath120 satisfies @xmath121 for @xmath122 . since the inequality ( [ l2-hellinger - inequal ] ) continues to hold in @xmath0 for @xmath5 ( with @xmath123 replaced by @xmath124 on the right side ) , we obtain the following corollary : [ l2-ratecorollarymultcase ] under the conditions of theorem [ newmulthellingerratethm ] it follows that @xmath125 for @xmath126 .
here we give the proof of theorem [ newmulthellingerratethm ] .
the main tool is a method developed by @xcite .
we will use the following lemma in combination with theorem 7.6 of @xcite or theorem 3.4.1 of @xcite ( section 3.4.2 , pages 330 - 331 ) . without loss of generality
we can take @xmath127 where @xmath128 is the upper bound of the support of @xmath22 ( see theorem [ newmulthellingerratethm ] ) .
let @xmath96 be a collection of probability densities @xmath129 on a sample space @xmath130 with respect to a dominating measure @xmath28 .
define @xmath131 } : \
p \in { \cal p } \right \ } , \qquad \mbox{for } \ \
\sigma > 0 .
\label{calgsubsigmadefn}\end{aligned}\ ] ] the following general result relating the bracketing entropies @xmath132 } ( \cdot , { \cal g}^{(conv ) } , l_2 ( p_0))$ ] , @xmath132 } ( \cdot , { \cal g}_{\sigma(\epsilon)}^{(conv ) } , l_2 ( p_0))$ ] , @xmath132 } ( \cdot , { \cal p } _ , l_2 ( q_{\sigma(\epsilon ) } ) ) $ ] , and @xmath132 } ( \cdot , { \cal p } _ , l_2 ( \tilde{q}_{\sigma(\epsilon ) } ) ) $ ] is due to @xcite .
( van de geer , 2000 ) [ vdgbasiclemma ] for every @xmath133 @xmath134 } ( 3 \epsilon , { \cal g}^{(conv ) } , l_2 ( p_0 ) ) & \le & \log n_{[\ , ] } ( \epsilon , { \cal g}_{\sigma(\epsilon)}^{(conv ) } , l_2 ( p_0 ) ) \label{vdg_first_inequality } \\ & \le & \log n_{[\ , ] } ( \epsilon/2 , { \cal p } , l_2 ( q_{\sigma(\epsilon ) } ) ) \label{vdg_second_inequality } \\ & = & \log n_{[\ , ] } \left ( \frac{\epsilon/2}{\sqrt{q_{\sigma(\epsilon ) } ( { \cal x } ) } } , { \cal p } , l_2 ( \tilde{q}_{\sigma(\epsilon ) } ) \right ) \label{vdg_third_equality } \end{aligned}\ ] ] where @xmath135 } d \mu$ ] and @xmath136 .
we first show that ( [ vdg_first_inequality ] ) holds .
suppose that @xmath137 , \
j = 1 , \ldots , m \}$ ] are @xmath138-brackets with respect to @xmath139 for @xmath140 with @xmath141 , \qquad m = n_{[\ , ] } ( \epsilon , { \cal g}_{\sigma(\epsilon)}^{(conv ) } , l_2 ( p_0 ) ) .\ ] ] then for @xmath142 , let @xmath143}$ ] be the corresponding element of @xmath140 .
suppose that @xmath144 $ ] for some @xmath145 . then @xmath146 } + g_{\sigma } \ \left \ { \begin{array}{l } \le g 1_{[p_0 \le \sigma ] } + g_{u , j } \equiv \tilde{g}_{u , j } \\ \ge 0 + g_{l ,
j } \equiv \tilde{g}_{l , j } , \end{array } \right .\end{aligned}\ ] ] where , by the triangle inequality , @xmath147 for all @xmath142 , and the definition of @xmath148 , it follows that @xmath149 thus @xmath150 : \ j \in \ { 1 , \ldots , m\ } \}$ ] is a collection of @xmath151brackets for @xmath152 with respect to @xmath153 and hence ( [ vdg_first_inequality ] ) holds .
now we show that ( [ vdg_second_inequality ] ) holds .
suppose that @xmath154 : \ j = 1 , \ldots , m \}$ ] is a set of @xmath155brackets with respect to @xmath156 for @xmath96 with @xmath157 \qquad \mbox{and } \qquad m = n_{[\ , ] } ( \epsilon/2 , { \cal p } , l_2 ( q_{\sigma(\epsilon ) } ) ) .\ ] ] suppose @xmath158 $ ] for some @xmath159 .
then , since @xmath160 } \ \ \left \ { \begin{array}{l } \le \frac{2p_{u , j}}{p_{u , j } + p_0 } 1_{[p_0 > \sigma ] } \equiv g_{u , j } , \\ \ \
\\ \ge \frac{2p_{l , j}}{p_{u , j } + p_0 } 1_{[p_0 > \sigma ] } \equiv g_{l , j } \end{array } \right .\end{aligned}\ ] ] where @xmath161 } - \frac{2p_{l , j}}{p_{u , j } + p_0 } 1_{[p_0 > \sigma ] } \bigg | \\ & = & \frac{2(p_{u , j } - p_{l , j})}{p_{l , j } + p_0 } 1_{[p_0 > \sigma ] } \le \frac{2 | p_{u , j } - p_{l , j } |}{p_0 } 1_{[p_0 > \sigma ] } .
\end{aligned}\ ] ] thus @xmath162 and hence @xmath163 : \ j = 1 , \ldots , m \}$ ] is a set of @xmath138-brackets with respect to @xmath153 for @xmath164 .
this shows that ( [ vdg_second_inequality ] ) holds .
it remains only to show that ( [ vdg_third_equality ] ) holds .
but this is easy since @xmath165 .
this lemma is based on @xcite , pages 101 and 103 .
note that our constants differ slightly from those of van de geer .
[ boundsforfzerof ] suppose that @xmath19 has density @xmath115 which satisfies , for some @xmath166 , @xmath167^d .
\label{boundsforfzerodensity}\end{aligned}\ ] ] then @xmath168 ( which we can identify with the vector @xmath169 ) satisfies @xmath170^d , \\ & & \vdots \\ & & p_{0,2^d } ( \underline{t } ; f_0 ) \left \ { \begin{array}{l } \le c_1 \prod_{j=1}^d ( 1-t_j ) \\
\ge c_1^{-1 } \prod_{j=1}^d ( 1-t_j ) , \end{array } \right .
\qquad \mbox{for all } \
\ \underline{t } \in [ 0,1]^d .\end{aligned}\ ] ] this follows immediately from the general @xmath171 version of ( [ structureofthecellprobabilities ] ) and the assumption on @xmath115 .
these inequalities can also be written in the following compact form : for @xmath172 with @xmath173 , @xmath174^d .\end{aligned}\ ] ] [ boundsforsmallvaluesdensity ] suppose that the assumption of lemma [ boundsforfzerof ] holds .
suppose , moreover , that @xmath26 has density @xmath27 which satisfies @xmath175^d .
\label{boundsforgzerodensity } \end{aligned}\ ] ] then @xmath176 } p_0 d \mu \le 2^d ( c_1 c_2)^2 \sigma .\end{aligned}\ ] ] furthermore , with @xmath177 we have @xmath178 } p_0 d \mu \le \delta^2 .\ ] ] the first inequality follows easily from lemma [ boundsforfzerof ] : note that @xmath179 } p_0 d\mu & = & \sum_{k=1}^{2^d } \int_{[p_k ( \underline{t } , f_0 ) \le \sigma ] } p_k ( \underline{t } , f_0 ) g_0 ( \underline{t } ) \underline{dt } \\
& \le & 2^d \int_{[f_0 ( \underline{t})g_0 ( \underline{t } ) \le \sigma ] } f_0 ( \underline{t } ) g_0 ( \underline{t } ) \underline{dt } \\ & \le & 2^d c_1 c_2 \int _ { [ c_1^{-1 } c_2^{-1 } \prod_{j=1}^d t_j \le \sigma ] } \prod_{j=1}^d t_j \ , \underline{dt } \le 2^d ( c_1 c_2)^2 \sigma .\end{aligned}\ ] ] the second inequality follows from the first inequality of the lemma .
[ massofqsigma ] if the hypotheses of lemmas [ boundsforfzerof ] and [ boundsforsmallvaluesdensity ] hold , then the measure @xmath180 defined by @xmath181 has total mass @xmath182 given by @xmath183^d : \
\prod_{j=1}^d t_j > \sigma/(c_1
c_2 ) \ } } \frac { c_1c_2 } { \prod_{j=1}^d t_j } \underline{dt } \label{firstcrackanalyticexpressiontotalmass } \\ & = & \frac{2^d c_1 c_2}{d ! } ( \log ( c_1 c_2 / \sigma ) ) ^d .
\label{asympapproxtotalmass}\end{aligned}\ ] ] this follows from lemma [ boundsforfzerof ] , followed by an explicit calculation . in particular , the equality in ( [ asympapproxtotalmass ] ) follows from @xmath184 } \frac{1}{\prod_{j=1}^d t_j } \underline{dt } & = & \int _ { [ \sum_1^d x_j \le \log ( 1/b ) ] } \underline{dx } \ \
\mbox{by the change of variables } \
t_j = e^{-x_j},\\ & = & \frac{1}{d ! } \left ( \log ( 1/b ) \right ) ^d \ \ \ \ \mbox{for } \ \ 0
< b \le 1\end{aligned}\ ] ] where the second equality follows by induction : it holds easily for @xmath7 ( and @xmath103 ) ; and then an easy calculation shows that it holds for @xmath171 if it holds for @xmath185 .
[ cruxentropybound ] if the hypotheses of lemmas [ boundsforfzerof ] and [ boundsforsmallvaluesdensity ] hold , and @xmath186 , then @xmath134 } ( \epsilon , { \cal g}^{(conv ) } , l_2 ( p_0 ) ) \le k \frac { \left [ \log ( 1/\epsilon)\right ] ^{5d/2 - 2}}{\epsilon } \end{aligned}\ ] ] for all @xmath187 and some constant @xmath188 .
this follows by combining the results of lemmas [ boundsforsmallvaluesdensity ] and [ massofqsigma ] with lemma [ vdgbasiclemma ] , and then using corollary [ corollariesofgaosthm ] of the bracketing entropy bound of @xcite and stated here as theorem [ gaosthm ] .
here is the explicit calculation : @xmath189 } ( 6 \epsilon , { \cal g}^{(conv ) } , l_2 ( p ) ) } \\ & \le & \log n_{[\ , ] } \left ( \frac{\epsilon}{\sqrt{q_{\sigma(\epsilon ) } ( { \cal x } ) } } , { \cal p } , l_2 ( \tilde{q}_{\sigma(\epsilon ) } ) \right ) \qquad \mbox{by lemma~\ref{vdgbasiclemma } } \\ & \le & \log n_{[\ , ] } \left ( \frac{\epsilon}{\sqrt{\frac{2^d c_1c_2}{d ! } \left [ \log ( ( c_1c_2)^3\cdot 2^d /(\epsilon^2 ) \right ] ^d } } , { \cal p } , l_2 ( \tilde{q}_{\sigma(\epsilon ) } ) \right ) \qquad \mbox{by \ lemmas~\ref{boundsforsmallvaluesdensity}~and~\ref{massofqsigma } } \\ & \le & \log n_{[\ , ] } \left ( \frac{v \epsilon}{\left [ \log ( 1/\epsilon ) \right ] ^{d/2 } } , { \cal p } , l_2 ( \tilde{q}_{\sigma(\epsilon ) } ) \right ) \qquad \mbox{for } \ \ v = v_d ( c_1,c_2 ) \\ & \le & k \frac{\left [ \log(1/\epsilon ) \right ] ^{d/2}}{v \epsilon } \left [ \log \left ( \frac { ( \log ( 1/\epsilon))^{d/2}}{v\epsilon } \right ) \right ] ^{2(d-1 ) } \qquad \mbox{by corollary~\ref{corollariesofgaosthm}(b ) } \\ & \le & \tilde{k } \frac{\left [ \log ( 1/\epsilon ) \right ] ^{5d/2 -2}}{\epsilon } \end{aligned}\ ] ] for @xmath138 sufficiently small .
( theorem [ newmulthellingerratethm ] ) this follows from lemma [ cruxentropybound ] and theorem 7.6 of @xcite or theorem 3.4.1 of @xcite together with the arguments given in section 3.4.2 . by lemma [ cruxentropybound ] the bracketing entropy integrals @xmath190 } ( \delta , { \cal g}^{(conv ) } , l_2 ( p_0 ) ) \equiv \int_0^{\delta } \sqrt { 1 + \log n_{[\ , ] } ( \epsilon , { \cal g}^{(conv ) } , l_2 ( p_0 ) ) } \
d\epsilon \lesssim \int_0^\delta \epsilon^{-1/2 } \left \ { \log ( 1/\epsilon ) \right \}^{3\gamma_d/2 } \ d \epsilon\end{aligned}\ ] ] where the bound on the right side behaves asymptotically as a constant times @xmath191 with @xmath192 , and hence ( using the notation of theorem 3.4.1 of @xcite ) , we can take @xmath193 .
thus with @xmath194 with @xmath195 we find that @xmath196 and hence the claimed order of convergence holds .
there are several related models in which we expect to see the same basic phenomenon as established here , namely a global convergence rate of the form @xmath1 in all dimensions @xmath5 with only the power @xmath197 of the log term depending on @xmath171 .
three such models are : + ( a ) the `` in - out model '' for interval censoring in @xmath0 ; + ( b ) the `` case 2 '' multivariate interval censoring models studied by @xcite ; and + ( c ) the scale mixture of uniforms model for decreasing densities in @xmath70 .
+ here we briefly sketch why we expect the same phenomenon to hold in these three cases , even though we do not yet know pointwise convergence rates in any of these cases . the `` in - out model '' for interval censoring in @xmath0 was explored in the case @xmath103 by @xcite . in this model @xmath198 on @xmath199
, @xmath200 is a random rectangle in @xmath199 independent of @xmath71 ( say @xmath201 = \ { \underline{x } = ( x_1 , x_2 ) \in { { \mathbb r}}^2 : \ u_1 \le x_1 \le v_1 , \
u_2 \le x_2 \le v_2 \}$ ] where @xmath202 and @xmath203 are random vectors in @xmath199 with @xmath204 coordinatewise ) .
we observe only @xmath205 , and the goal is to estimate the unknown distribution function @xmath22 .
@xcite ( page 86 ) produced a local asymptotic minimax lower bound for estimation of @xmath22 at a fixed @xmath206 . under the assumption that @xmath22 has a positive density @xmath207 at @xmath112
, @xcite showed that any estimator of @xmath208 can have a local - minimax convergence rate which is at best @xmath6 .
@xcite has shown that this rate can be achieved by estimators involving smoothing methods .
based on the results for current status data in @xmath0 obtained in theorem [ newmulthellingerratethm ] and the entropy results for the class of distribution functions on @xmath0 , we conjecture that the global hellinger rate of convergence of the mle @xmath209 will be @xmath210 for all @xmath211 where @xmath212 .
recall that `` case 2 '' interval censored data on @xmath10 is as follows : suppose that @xmath213 on @xmath12 , the pair of observation times @xmath214 with @xmath215 determines a random interval @xmath216 $ ] , and we observe @xmath217 where @xmath218 , @xmath219 , and @xmath220 .
nonparametric estimation of @xmath19 based on @xmath221 i.i.d . as @xmath89 has been discussed by a number of authors , including @xcite , @xcite , and @xcite .
@xcite studied generalizations of this model to @xmath0 , and obtained rates of convergence of the mle with respect to the hellinger metric given by @xmath222 in the case most comparable to the multivariate interval censoring model studied here . while this rate reduces when @xmath7 to the known rate @xmath223 , it is slower than @xmath210 for some @xmath224 when @xmath225 due to the use of entropy bounds involving convex hulls ( see @xcite , proposition a.1 , page 66 ) which are not necessarily sharp .
we expect that rates of the form @xmath210 with @xmath226 are possible in these models as well .
@xcite and @xcite studied the family of scale mixtures of uniform densities of the following form : @xmath227 } ( \underline{x } ) dg(\underline{y})\equiv \int_{{{\mathbb r}}^{+d } } \frac{1}{|\underline{y}| } 1_{(0,\underline{y } ] } ( \underline{x } ) dg(\underline{y } ) \label{scalemixtureofuniforms}\end{aligned}\ ] ] for some distribution function @xmath49 on @xmath228 .
( note that we have used the notation @xmath229 for @xmath230 . )
it is not difficult to see that such densities are decreasing in each coordinate and that they also satisfy @xmath231 = ( -1)^d \int_{(\underline{u},\underline{v } ] } |\underline{y}|^{-1 } 1_{(\underline{y},\underline{v } ] } d g(\underline{y } ) \ge 0\end{aligned}\ ] ] for all @xmath232 with @xmath233 ; here @xmath234 denotes the @xmath9dimensional difference operator .
this is the same key property of distribution functions which results in ( bracketing ) entropies which depend on dimension only through a logarithmic term .
the difference here is that the density functions @xmath235 need not be bounded , and even if the true density @xmath236 is in this class and satisfies @xmath237 , then we do not yet know the behavior of the mle @xmath238 at zero .
in fact we conjecture that : ( a ) if @xmath237 and @xmath115 is a scale mixture of uniform densities on rectangles as in ( [ scalemixtureofuniforms ] ) , then @xmath239 for some @xmath240 .
( b ) under the same hypothesis as in ( a ) and the hypothesis that @xmath115 has support contained in a compact set , the mle converges with respect to the hellinger distance with a rate that is no worse than @xmath241 where @xmath242 . again
@xcite and @xcite establish asymptotic minimax lower bounds for estimation of @xmath243 proving that no estimator can have a ( local minimax ) rate of convergence faster than @xmath6 in all dimensions .
this is in sharp contrast to the class of block - decreasing densities on @xmath70 studied by @xcite and by @xcite : @xcite shows that the local asymptotic minimax rate for estimation of @xmath244 is no faster than @xmath245 , while @xcite show that there exist ( histogram type ) estimators @xmath246 which satisfy @xmath247 .
we begin by summarizing the results of @xcite . for a ( probability )
measure @xmath28 on @xmath248^d$ ] , let @xmath249 denote the corresponding distribution function given by @xmath250 ) = \mu ( [ 0,x_1]\times \cdots \times [ 0,x_d ] ) \ ] ] for all @xmath251^d$ ] .
let @xmath252 denote the collection of all distribution functions on @xmath248^d$ ] ; i.e. @xmath253^d \ } .\end{aligned}\ ] ] for example , if @xmath254 denotes lebesgue measure on @xmath248^d$ ] , then the corresponding distribution function is @xmath255 .
( gao , 2012 ) .
[ gaosthm ] for @xmath211 and @xmath256 @xmath257 } ( \epsilon , { \cal f}_d , l_p ( \lambda_d ) ) \lesssim \epsilon^{-1 } \left ( \log ( 1/\epsilon ) \right ) ^{2(d-1 ) } \end{aligned}\ ] ] for all @xmath258 .
our goal here is to use this result to control bracketing numbers for @xmath252 with respect to two other measures @xmath259 and @xmath260 defined as follows .
let @xmath259 denote the finite measure on @xmath248^d$ ] with density with respect to @xmath254 given by @xmath261 for fixed @xmath262 , let @xmath260 denote the ( probability ) measure on @xmath263^d$ ] with density with respect to @xmath254 given by @xmath264 [ corollariesofgaosthm ] ( a ) for each @xmath211 it follows that for @xmath265 @xmath134 } ( 2^{d/2}\epsilon , { \cal f}_d , l_2 ( c_d ) ) \lesssim \epsilon^{-1 } \left ( \log ( 1/\epsilon ) \right ) ^{2(d-1 ) } .\end{aligned}\ ] ] ( b ) for each @xmath211 and @xmath266 it follows that for @xmath267 @xmath134 } ( 2^{d/2 + 1}\epsilon , { \cal f}_d , l_2 ( r_{d,\sigma } ) ) \lesssim \epsilon^{-1 } \left ( \log ( 1/\epsilon ) \right ) ^{2(d-1 ) } .\end{aligned}\ ] ] we first prove ( a ) .
we set @xmath268 where @xmath269 and @xmath270 satisfy @xmath271 .
let @xmath272 , j = 1 , \ldots , m\}$ ] be a collection of @xmath273brackets for @xmath252 with respect to @xmath274 .
( thus for @xmath103 , @xmath275 , @xmath276 , and @xmath277 , while for @xmath278 , @xmath279 , @xmath280 , and @xmath281 . ) by theorem a.1 we know that @xmath282 .
now we bound the size of the brackets @xmath283 $ ] with respect to @xmath259 . using hlder s inequality with @xmath284
as chosen above we find that @xmath285^d } ( h_j - g_j ) ^2 c_d ( u ) du & \le & \left ( \int_{[0,1]^d } | h_j - g_j |^{2r } \underline{du } \right ) ^{1/r } \cdot \left ( \int_{[0,1]^d } c_d ( \underline{u})^s \underline{du } \right ) ^{1/s } \nonumber \\ & \le & ( \epsilon^p ) ^{1/r } \cdot 2^{d / s } \le 2^{d } \epsilon^2 .
\label{l2entropywrtcd}\end{aligned}\ ] ] here are some details of the computation leading to ( [ l2entropywrtcd ] ) : @xmath285^d } c_d ( \underline{u})^s \underline{du } & = & \int_{[0,1]^d } \left ( \frac{d!}{d^d } \right ) ^s \prod_{j=1}^d \frac{1}{u_j^{(d-1/2)/d } } \cdot 1 \left \ { \sum_{j=1}^d u_j^{1/d } > d-1 \right \ } \underline{du } \\ & = & \left ( \frac{d!}{d^d } \right ) ^s \cdot ( 2d)^d \int_{[0,1]^d } 1 \left \ { \sum_{j=1}^d x_j^{2 } > d-1 \right \ } \underline{dx}\\ & \le & \left ( \frac{d!}{d^d } \right ) ^s \cdot ( 2d)^d \cdot \int_{[0,1]^d}1 \left \ { \sum_{j=1}^d x_j
d-1 \right \ } \underline{dx}\\ & \le & \left ( \frac{d!}{d^d } \right ) ^s \cdot ( 2d)^d \cdot \int_{[0,1]^d}1 \left \ { \sum_{j=1}^d t_j < 1 \right \ } \underline{dt}\\ & = & 2^d\left ( \frac{d!}{d^d } \right ) ^{s-1 } \le 2^d .\end{aligned}\ ] ] to prove ( b ) we introduce monotone transformations @xmath286 and their inverses @xmath287 which relate @xmath288 and @xmath289 : we set @xmath290 for @xmath291 .
these all depend on @xmath292 , but this dependence is suppressed in the notation . for the same brackets
@xmath293 $ ] used in the proof of ( a ) , we define new brackets @xmath294 $ ] for @xmath291 by @xmath295 then it follows easily by direct calculation using @xmath296 that @xmath285^d } ( \tilde{h}_j ( t ) - \tilde{g}_j ( t))^2 r_{d , \sigma } ( t ) \underline{dt } & = & \int_{[0,1]^d } ( h_j ( u ) - g_j ( u))^2 c_d ( u ) \underline{du } .\end{aligned}\ ] ] thus for @xmath297 we have @xmath298 by the arguments in ( a ) .
hence the brackets @xmath299 $ ] yield a collection of @xmath300 brackets for @xmath252 with respect to @xmath301 , and this implies that ( b ) holds . | we establish global rates of convergence of the maximum likelihood estimator ( mle ) of a multivariate distribution function on @xmath0 in the case of ( one type of ) `` interval censored '' data .
the main finding is that the rate of convergence of the mle in the hellinger metric is no worse than @xmath1 for @xmath2 . |
we first present our results in the absence of applied electric fields .
figure [ fig : phasediagram ] shows the calculated temperature - disorder phase diagram and the zero temperature free energies where we have identified three regions according to the rf strength . for weak rfs ( @xmath42 ) ,
long - range fe order sets in at a transition temperature @xmath43 and it is accompanied by a metastable random field disordered state down to @xmath44 , as it is shown in fig.[fig : phasediagram](a ) . for moderate compositional disorder ( @xmath45 ) , there is no transition as the the rf state becomes stable at all temperatures and the long - ranged polar state is now metastable , see fig.[fig : phasediagram](b ) . for strong compositional disorder ( @xmath46 ) , only the rf state exists , as it is shown in fig.[fig : phasediagram](c ) . by comparing our phase diagram with that of the relaxor pbmg@xmath0nb@xmath1o@xmath2-pbtio@xmath2 ( pmn - pt ) @xcite and assuming that the conventional fe pto is near about @xmath36 with @xmath47 , then pmn is in the weak - disorder region with @xmath48 and @xmath49 , as it is shown in fig.[fig : phasediagram](a ) .
this means that while the ground state of pmn - pt relaxors is fe , those in the ti - poor side of the morphotropic phase boundary are stuck in a metastable disordered random field state below the phase transition line .
we will see below that this is also supported by the predicted correlation lengths and static susceptibilities of our model . the temperature and disorder dependence of the zone - center to phonon frequency , @xmath50 , and the order parameter , @xmath51 , associated with the rf and fe states are shown in figure [ fig : polarization ] ( a ) and ( b ) , respectively . while the to mode softens and condenses at @xmath52 for the pure case ,
as expected , that of the rf state remains finite all the way down to zero temperature for any amount of compositional disorder .
the temperature dependence of the metastable states is shown for the sake of completeness . .
( * b*)-(*c * ) zero temperature free energies for moderate ( @xmath53 and strong ( @xmath54 ) compositional disorder , respectively , showing the rf state as a global minimum .
@xmath55 is the order parameter for the pure case at zero temperature .
all energies are plotted with respect to their corresponding minimum . ] . here ,
@xmath56 and @xmath57 and @xmath58 are the to frequency and order parameter of the pure fe at zero temperature . ]
we now discuss the correlation functions in the fluctuations of polarization .
as it is usually done for conventional cubic fes @xcite , we consider mean squared fluctuations on the polarization components @xmath59 and @xmath60 that are transverse and longitudinal to a wave - vector @xmath20 , respectively .
for the transverse components we obtain isotropic fluctuations with the following form , @xmath61 where we have identified @xmath62 as the correlation length .
@xmath63 denotes thermal and compositional averages taken in that order . in the absence of disorder and in the classical limit ( @xmath64 ) ,
( [ eq : sqtransverse ] ) reproduces to the fluctuations of pure fes .
@xcite in the classical limit , eq .
( [ eq : sqtransverse ] ) becomes a lorentzian plus a lorentzian squared .
while this is analogous to the well - known result of the random field ising model @xcite , we will show below that the correlation functions behave very differently in real space due to the anisotropy and long - range nature of the dipole force .
the wave - vector distribution predicted by eq .
( [ eq : sqtransverse ] ) has been recently observed in diffuse scattering experiments @xcite and the quantum fluctuations have been found important to correctly describe the observed static structure factor at low temperatures @xcite .
for the fluctuations in @xmath60 , we find that they have a similar form to that of eq .
( [ eq : sqtransverse ] ) except that the to frequency is replaced by that of the the longitudinal mode @xmath65 , where @xmath66 is a constant related to the depolarizing field @xcite , which makes these fluctuations significantly smaller than those in @xmath59 .
figure [ fig : correlationlength ] shows the calculated temperature dependence of the correlation length @xmath67 of the rf state for several disorder strengths . in the absence of disorder , @xmath67 diverges as expected near the fe transition . in the presence of disorder ,
the correlation length of the random field state remains finite at all temperatures .
at @xmath44 it scales with disorder as @xmath68 , which we identify as the minimum length scale on which domains must appear spontaneously . by a standard procedure @xcite , it can be shown that the static dielectric susceptibility is given by @xmath69 .
the temperature and disorder dependence of the resulting static dielectric constant @xmath70 , are shown in the inset of fig .
[ fig : correlationlength ] .
we find that our model is in fair qualitative and quantitative agreement with the measured correlation length @xcite and static dielectric constant @xcite in pmn when @xmath71 .
this is consistent with our identification of the ti - poor region of pmn - pt in our phase diagram ( see fig . [
fig : phasediagram ] ) and where the rf state is metastable . .
] ) disorder ( @xmath72 ) at several temperatures . here , @xmath73 .
the inset in ( a ) is a log - log plot of the longitudinal correlations showing the crossover from exponential to power - law behavior ; and the insets in ( c ) and ( d ) show the change in sign in the correlations with increasing distance . ]
we now discuss the spatial dependence of the correlation functions of polarization @xmath74 .
figure [ fig : cfs ] shows @xmath75 calculated from fourier transform of eq .
( [ eq : sqtransverse ] ) . in the presence of compositional disorder ,
the correlations of the random field state are positive along the longitudinal direction ( @xmath76 ) and they change sign in the transverse direction ( @xmath77 ) , as it is shown in figs .
[ fig : cfs](a ) and [ fig : cfs](b ) , respectively . for short distances compared to the correlation length @xmath67 , they fall - off exponentially and then cross over to a power law behavior ( @xmath78 ) for @xmath79 , see inset in fig .
[ fig : cfs](a ) and sm .
note they also increase with decreasing temperature but do not reach long - range order as their correlation length @xmath67 remains finite for all temperatures .
this is in stark contrast with the correlations of the pure compound where , while anisotropic , they are strongest near the fe transition and then weaken away from it , as it is shown in figs . [
fig : cfs](c ) and [ fig : cfs](d ) .
the cross - component correlations ( @xmath80 ) of the random field state are identically zero , as expected from cubic symmetry . a previous theoretical work on the relaxor @xmath81%pmn-@xmath82%pt has also found anisotropic correlations @xcite .
we now describe our results in the presence of an applied electric field .
figure [ fig : pmnpolarization ] , shows the temperature dependence of the order parameter for weak disorder and several field strengths .
for weak applied electric fields @xmath83 , the polarization of the rf state grows with decreasing temperature without inducing a fe transition . for moderate field strengths @xmath84
, a clear first - order transition occurs as shown by the discontinuity in the order parameter . this discontinuity becomes weaker with increasing applied field until it reaches a critical point where the transition is of second order .
for strong applied fields @xmath85 , the transition is smeared .
this behaviour is in agreement with experiments in pmn - pt @xcite . , orange line ) . upon increasing the field strength ( @xmath86 , red line )
the system approaches a critical point , as observed in experiments.@xcite here @xmath87 and @xmath55 is the order parameter at zero temperature for the pure compound . ]
to summarize , we have studied the effects of cubic random electric fields on the lattice instabilities the lead a the fe transition .
we have shown that a rf state emerges from the combined effect of dipolar forces and compositional disorder .
such state has no - long range fe order and anisotropic , long - ranged correlations of polarization that grow with decreasing temperature .
when comparing to the experimental phase diagram of typical relaxors such as pmn - pt , we conclude that while the ground state is fe , the rf state in the poor ti side of the phase diagram is metastable below the phase transition line down to zero kelvin . upon application of strong enough electric fields , first - order transitions can be induced and end at a critical point .
while we have focused our attention on pmn - pt , our model and results are generic and should be applicable to other relaxors such as pbzn@xmath0nb@xmath1o@xmath2-pbtio@xmath2 .
* variational solution*. we now describe our variational solution of our model hamiltonian of eq .
( [ eq : hamiltonian ] ) .
we consider the trial probability distribution , @xmath88 where @xmath89 is the hamiltonian of a displaced , cubic harmonic oscillator in a random field @xmath30 , @xmath90 and @xmath91 its normalization , @xmath92^{-1 } \right ) \times e^{\frac{\beta}{2 } \sum _ { ij,\lambda \lambda^\prime } h_{i \lambda } ( \operatorname{\mathcal{d}}_{i - j}^{-1 } ) ^{\lambda \lambda^\prime } h_{j \lambda^\prime } + \beta \sum_{i \lambda } h_{i \lambda } a _
{ i \lambda } } , \end{aligned}\ ] ] where @xmath93 is a component of the dynamical matrix and @xmath94 are the to soft mode frequencies at wave vector @xmath20 and are given by the squared of the the eigenvalues of the fourier transform of the dynamical matrix @xmath95 .
@xmath96 is the @xmath26-component of the order parameter at site @xmath11 and it corresponds to the mean displacement averaged over thermal and compositional disorder , @xmath97 the mean squared fluctuations between local soft mode components @xmath98 and @xmath99 averaged over disorder are given as follows , @xmath100 where @xmath101 @xmath102 is a unitary transformation that takes @xmath103 to its diagonal representation .
it is convenient to define the following matrix , @xmath104 @xmath105 fourier transforms as follows , @xmath106 where , @xmath107 we now compute the free energy @xmath108 using our probability distribution ( [ eq : rhotrial ] ) together with the above equations .
the result is the following , @xmath109 + \frac{\gamma_1}{4 } \sum_{i \lambda } \left [ 3 \left ( \psi_{ii}^ { \lambda \lambda } \right)^2 + 6 \psi_{ii}^ { \lambda \lambda } a_{i \lambda}^2 + a _
{ i \lambda } ^4 \right ] \nonumber \\ & + \frac{\gamma_2}{4}\sum_{i , \lambda \neq \lambda^\prime } \left [ 2 \left ( \psi_{ii}^ { \lambda \lambda^\prime } \right)^2 + \psi_{ii}^ { \lambda \lambda } \psi_{ii}^ { \lambda^\prime \lambda^\prime } + \psi_{ii}^ { \lambda \lambda } a _ {
i \lambda^\prime } ^2 + 4 \psi_{ii}^ { \lambda \lambda^\prime } a _
{ i \lambda } a _ { i \lambda^\prime } + \psi_{ii}^ { \lambda^\prime \lambda^\prime } a _ {
i \lambda } ^2 + a _ {
i \lambda } ^2 a _
{ i \lambda^\prime } ^2 \right ] \\ & - \frac{1}{2 } \sum_{ij , \lambda \lambda^\prime } v_{ij}^ { \lambda \lambda^\prime } \left [ a_{i\lambda } a_{j \lambda^\prime } + \psi_{ij}^{\lambda \lambda^\prime } \right ] - \delta^2 \sum _ { i \lambda } ( \operatorname{\mathcal{d}}_{ii}^{-1})^ { \lambda \lambda } - \sum_{i \lambda } e_{i \lambda } ^0 a _ { i \lambda } \nonumber \\ & - k_b t n \sum _ { { \bm q } \lambda } \left\ { \beta \omega _ { { \bm q } \lambda } ^2 \tilde{\eta } _ { { \bm q } \lambda } - \ln\left [ 2 \sinh\left ( \frac { \beta \omega _ { { \bm q } \lambda } } { 2 } \right ) \right ] \right\ } , \nonumber\end{aligned}\ ] ] where , @xmath110 minimization of the free energy ( [ eq : free1 ] ) with respect to @xmath116 gives the following result , @xmath117 a_{\nu } = e_{\alpha}^0 a_\alpha.\end{aligned}\ ] ] in writing eq .
( [ eq : p ] ) we have used the property that the summation @xmath118 does not depend on the origin @xmath11 ( @xmath119 is translationally invariant ) .
@xmath120 depends on the direction in which @xmath121 because @xmath122 is non - analytic .
( [ eq : dm ] ) and ( [ eq : p ] ) are the starting point of our analysis , see sm .
.1 in * acknowledgments*. work at the university of costa rica is supported by the vice - rectory for research under the project no .
816-b5 - 220 .
work at argonne is supported by u.s .
department of energy , office of basic energy sciences science , materials sciences and engineering division under contract no .
de - ac02 - 06ch11357 .
gggv acknowledges useful discussions with peter b. littlewood .
43ifxundefined [ 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 10.1080/00150190490423057 [ * * , ( ) ] http://dx.doi.org/10.1063/1.365983 [ * * , ( ) ] http://stacks.iop.org/0953-8984/15/i=2/a=110 [ * * , ( ) ] link:\doibase 10.1080/01411594.2014.989229 [ * * , ( ) ] link:\doibase 10.1038/nmat3951 [ * * , ( ) ] link:\doibase 10.1126/science.1129564 [ * * , ( ) ] link:\doibase 10.1111/j.1151 - 2916.1954.tb13992.x [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.60.13470 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.137601 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.144102 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.147602 [ * * , ( ) ] link:\doibase 10.1103/physrevb.91.214117 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.116.127601 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.017601 [ * * , ( ) ] link:\doibase 10.1107/s0108767311040281 [ * * , ( ) ] link:\doibase 10.1073/pnas.1314780111 [ * * , ( ) ] link:\doibase 10.1080/01411594.2014.983508 [ * * , ( ) ] link:\doibase 10.1080/01411594.2014.989226 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.117.155501 [ * * , ( ) ] http://dx.doi.org/10.1038/ncomms4683 http://10.1038/ncomms4683 http://www.nature.com/articles/ncomms4683{#}supplementary-information [ * * , ( ) ] http://advances.sciencemag.org/content/2/9/e1501814 [ * * , ( ) ] http://dx.doi.org/10.1038/ncomms11014 http://10.1038/ncomms11014 http://www.nature.com/articles/ncomms11014{#}supplementary-information [ * * , ( ) ] link:\doibase 10.1080/00018732.2011.555385 [ * * , ( ) ] link:\doibase 10.1126/science.1129564 [ * * , ( ) ] http://stacks.iop.org/0953-8984/17/i=27/a=010 [ * * , ( ) ] link:\doibase 10.1103/physrevb.76.064122 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2963474 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.224109 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.68.847 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.35.1399 [ * * , ( ) ] @noop _ _ ( , , ) @noop _ _ ( , , ) link:\doibase 10.1038/nature04854 [ * * , ( ) ] link:\doibase 10.1103/physrevb.5.3758 [ * * , ( ) ] link:\doibase 10.1103/physrevb.8.3323 [ * * , ( ) ] link:\doibase 10.1103/physrevb.91.144105 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.134106 [ * * , ( ) ] link:\doibase 10.1103/physrev.187.1077 [ * * , ( ) ] @noop _ _ ( , , ) @noop _ _ ( , , ) @noop * * , ( ) link:\doibase 10.1103/physrevb.57.11204 [ * * , ( ) ] * supplemental material - random electric field instabilities of relaxor ferroelectrics *
for the cubic phase , there is no long - range order ( @xmath123 ) and , by symmetry , @xmath124 @xcite . therefore , the dynamical matrix has the form , @xmath125 \delta _ { \lambda \lambda^\prime } + \left ( v_{\bm q } \right ) _ { \lambda \lambda^\prime}.\end{aligned}\ ] ] for an arbitrary direction of @xmath20 the diagonalization of @xmath126 gives a doubly degenerate transverse optic ( to ) mode @xmath127 , and a singlet longitudinal optic ( lo ) mode @xmath128 , given as follows , [ eq : toloc ] @xmath129 where , @xmath130 is the zone - center to mode frequency and @xmath131 . for cubic symmetry ,
the transformation matrix @xmath132 that diagonalizes the dynamical matrix ( [ eq : dmcubic ] ) takes the form , @xmath133 where @xmath134 and @xmath135 are the usual azimuthal and polar angles in spherical coordinates .
we now calculate the fluctuations @xmath136 , @xmath137 where @xmath138 and , [ eq : psi ] @xmath139 we replace the summations by integrals over a sphere of wave - vector @xmath140 , @xmath141 by calculating the angular integrals , we find the result , @xmath142 thus , @xmath143 equations ( [ eq : toloc ] ) , ( [ eq : toloc0 ] ) , ( [ eq : psi ] ) , ( [ eq : psi0 ] ) determine de temperature and disorder dependence of the to and lo mode frequencies for the cubic phase .
for the rhombohedral phase , we assume a homogenous order parameter along the cube diagonal , @xmath144 .
also , by symmetry , @xmath145 @xcite .
the dynamical matrix is as follows , @xmath146 we first identify the soft mode frequencies .
pure longitudinal and transverse modes are obtained for wavevectors in the @xmath147 direction and the plane transverse to it . for @xmath148 diagonalization of the dynamical matrix
gives two distinct to mode frequencies @xmath149 and @xmath150 and one lo frequency @xmath151 at the zone - center .
for @xmath152 there is a doubly degenerate to mode frequency @xmath153 and one lo mode frequency @xmath154 at the zone - center . @xmath155 and @xmath156
are given as follows , [ eq : omegarhomboq1 ] @xmath157 while exact expressions can be derived for the phonon dispersions from the dynamical matrix @xmath158 , they are too elaborated and not enlightening
. instead we calculate them from perturbation theory .
our unperturbed basis is that of the cubic phase , therefore making the frequency splitting @xmath159 the expansion parameter .
it is also convenient to write the wavevector as @xmath160 , where @xmath161 is a right - handed coordinate system where @xmath162 is along the @xmath147 direction and @xmath163 are transverse to it .
the result is as follows , [ eq : omegarhomboq2 ] @xmath164 \frac { q_{t}^2 } { \left| { \bm q } \right|^2 } , \\ \left ( \omega _ { { \bm q } 3 } \right)^2 & = \left ( \omega_{0 1}^{\perp } \right)^2 + a_4 \left| { \bm q } \right|^2 + 3 a_3 + \left [ \left ( \omega_{0 3}^{\perp } \right)^2 - \left ( \omega_{0 1}^{\perp } \right)^2 \right ] \frac { q_l^2 } { \left| { \bm q } \right|^2 } , \end{aligned}\ ] ] where @xmath165 and with a transformation matrix given by , @xmath166 for an applied field @xmath167 , minimization of the free energy with respecto to the order parameter gives the following result , @xmath168 a = e_0,\end{aligned}\ ] ] which can be rewritten in terms of the soft mode frequencies as follows , @xmath169 a = e_0.\end{aligned}\ ] ] we now calculate @xmath170 and @xmath171 .
@xmath172 where [ eq : psirq ] @xmath173 to proceed further , we replace the summations by integrals over a sphere of wavevector @xmath140 , as we did in in the previous section and calculate the integrals over @xmath135 .
the result is the following , [ eq : psir0 ] @xmath174 \cos 2 \theta ) .
\end{aligned}\ ] ] equations ( [ eq : omegarhomboq1])-([eq : psir0 ] ) determine de temperature and disorder dependence of the order parameter @xmath51 and the to and lo mode frequencies in the rhombohedral phase .
in this section we derive analytic expressions for the correlation functions of the random field state .
the space dependence of the correlation functions of polarization @xmath175 , is given by the fourier transform of the fluctuations of polarization @xmath176 given in eq .
( [ eq : psiq ] ) . in the classical limit and in the cubic phase
, we calculate @xmath177 for distances larger than the correlation length @xmath67 , by taking replacing the summations by integrals .
we obtain the following results , @xmath178 where @xmath179 is the volume of the brillouin zone .
as stated in the main text , these correlations decrease as power laws for large distances ( @xmath180 ) and are anisotropic .
note that the corrections to these power laws are exponentially small .
table [ t : parameters ] gives the model parameters used in this work , as explained in the main text .
1ifxundefined [ 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.5.3758 [ * * , ( ) ] | relaxor ferroelectrics are complex oxide materials which are rather unique to study the effects of random field disorder on phase transitions . unlike the mostly studied random field magnets where the order parameter is uniaxial or isotropic , the polarization of typical relaxors lives in a cubic environment .
moreover , the exchange interaction that drives the magnetic transition is short - ranged and isotropic , while the relevant interaction in ferroelectrics is the highly anisotropic and long - ranged dipolar force .
this puts relaxors in a different universality class from that of disordered magnets making the standard model of random field disorder inadequate to describe their unusual properties . here , we study the effects of cubic random electric fields on the lattice instabilities that lead to the ferroelectric transition and show that , within a microscopic model and a statistical mechanical solution , even weak compositional disorder can prohibit the development of long - range order and that a random field state with anisotropic and power - law correlations of polarization emerges from the combined effect of dipole forces and their inherent charge disorder .
we compare and reproduce several key experimental observations in the well - studied relaxor pbmg@xmath0nb@xmath1o@xmath2-pbtio@xmath2 .
relaxors exhibit a myriad of complex phenomena that are both scientifically interesting and technologically important such as diffuse phase transitions where large and frequency dependent dielectric permittivities extend over hundreds of kelvin degrees @xcite without any signature of macroscopic symmetry breaking as well as unparalleled electromechanical responses @xcite .
these properties make relaxors attractive material candidates for energy storage and harvesting applications as well as future cooling technologies for integrated microelectronics @xcite .
though relaxors were first synthesised in the 1950s@xcite and they have been the subject of many theoretical @xcite and experimental studies @xcite there is still no consensus on a satisfactory theory of relaxor ferroelectricity @xcite .
one of the major difficulties in describing relaxors is that they exhibit several special energy scales : for temperatures above the so - called burns temperature @xmath3 their dielectric constant follows the curie - weiss law with a characteristic curie - weiss temperature @xmath4 . below @xmath3
, it deviates from the curie - weiss behavior and reaches a broad , frequency dependent maximum at temperature @xmath5 without any signature of a macroscopic broken symmetry . at low temperatures ,
no macroscopic structural changes are observed unless large enough electric fields are applied , which reveal an additional energy scale @xmath6 .
significant progress has been made by neutron and x - ray scattering experiments @xcite , which have observed the onset of anisotropic , quasi - elastic diffuse scattering at a temperature @xmath7 near about @xmath4 .
it has been recognized that a central question in the discussion of relaxors is the effect of random electric fields on the ferroelectric ( fe ) transition of cubic systems such as the typical perovskite relaxor pbmg@xmath0nb@xmath1o@xmath2 ( pmn ) @xcite .
the random electric fields originate from charge disorder : cations with different charge valencies are randomly located on the octahedrally coordinated site such as mg@xmath8 and nb@xmath9 in pmn .
@xcite these ions do not order with temperature , making the compositional disorder quenched .
unlike the widely studied random fields ( rfs ) in magnets which linearly couple to an order parameter of the ising or heisenberg type@xcite , the quenched electric rfs of relaxors couple to a cubic order parameter @xcite .
it is believed that @xmath7 is the onset temperature of a rf state in which relaxors exist @xcite .
in addition to the symmetry of the order parameter , we make the observation that the characteristic dipolar interaction of fes is equally important .
it is well - known that the structural instability that leads to the breaking of lattice inversion symmetry and a spontaneous polarization , is the result of dipolar forces between electric dipole moments induced by the displacements of the ions associated with a zone - center transverse optic ( to ) mode @xcite .
such dipolar forces are highly anisotropic and long - ranged , which are very much in contrast with the isotropic and short - ranged exchange couplings between the spin degrees of freedom of magnets . according to the theory of phase transitions @xcite , fes and magnets are therefore in different universality classes , rendering the standard models that describe the effects of rfs on magnetic transitions @xcite inadequate for relaxors @xcite .
the purpose of this work is then to study the random electric field problem posed by cubic relaxors within a minimal microscopic model .
we will show that as a result of the combined effect of dipolar forces and quenched rfs a state with no - long range fe order and anisotropic , long - ranged fluctuations of polarization emerges for any amount of compositional disorder .
we identify this disordered state as the rf state of relaxors .
we will also show that long - ranged fe order can be induced by application of strong enough electric fields and that such transition ends at a critical point , as it is observed in experiments.@xcite we consider a cubic lattice and choose normal mode coordinates that describe local displacements @xmath10 in the unit cell @xmath11 that are associated with the soft to mode , the condensation of which leads to the fe transition @xcite .
we consider the model hamiltonian , @xmath12 with @xmath13 .
@xmath14 is the conjugate momentum of @xmath15 ; @xmath16 is an applied electric field ; and @xmath17 is the dipolar interaction tensor with fourier transform @xmath18 \delta _ { \lambda \lambda^\prime } - 3 a_3 \frac { q _ { \lambda } q _ { \lambda^\prime } } { |{\bm q}|^2 } , $ ] where @xmath19 is the magnitude of the wavevector @xmath20 ; and @xmath21 and @xmath22 are constants that depend on the lattice structure @xcite .
hereafter , we denote @xmath23 as the component of @xmath24 when @xmath25 in the direction transverse to @xmath26 ( the value of @xmath27 depends on the direction in which @xmath20 approaches zero ) .
@xmath28 is the lattice stiffness and @xmath29 are anharmonic coefficients . for the quenched random fields @xmath30
, we choose a gaussian probability distribution of independent random variables with zero mean and variance @xmath31 . in the absence of compositional disorder , this is a standard minimal model for ferroelectricity in cubic perovskites .
@xcite to study the statistical mechanics of the hamiltonian ( [ eq : hamiltonian ] ) , it is necessary to consider thermal and quantum fluctuations at least at the level of the onsager approximation and random field fluctuations at least at the level of a replica theory .
@xcite to do so , we generalize a variational method previously developed by one of us @xcite to cubic symmetries .
such method allow us to calculate the temperature and disorder dependence of relevant quantities such as the phonon frequencies , the polarization order parameter and the correlation functions in a self - consistent fashion .
the details are presented in the methods section and supplementary material ( sm ) .
our model parameters are @xmath32 and @xmath33 . throughout this work ,
we have fixed the values of @xmath34 to those of typical values of oxide perovskites @xcite and to fit the transition temperature of the conventional fe pbtio@xmath2 ( pto , @xmath35k ) @xcite assuming @xmath36 .
the resulting values are given in the sm .
our choice gives a curie - weiss constant of @xmath37k and a zone - center to phonon energy of @xmath38mev at zero temperature , which are typical of conventional fes .
depending on the choice of the anharmonic coefficients , the low temperature fe phase predicted by the hamiltonian ( [ eq : hamiltonian ] ) in the absence of compositional disorder has tetragonal ( @xmath39 ) or rhombohedral ( @xmath40 ) symmetry @xcite . in this work ,
we have chosen @xmath41 , as we will study the field - induced fe transition of relaxors , which is typically a cubic - to - rhombohedral structural phase change . |
total colouring has attracted a considerable amount of attention since the infamous total colouring conjecture ( tcc ) was proposed by behzad @xcite and also by vizing @xcite .
the tcc asks whether every graph with maximum degree @xmath5 admits a @xmath6total colouring , namely an assignment of one of @xmath6 colours to each vertex and to each edge of the graph so that there is no monochromatic pair of adjacent vertices , pair of adjacent edges , or pair of incident vertex and edge .
we study the circular version of total colouring in this paper .
since a total colouring of a graph @xmath7 is in turn a vertex colouring of the total graph of @xmath7 , all basic properties of circular colourings carry over to circular total colouring .
circular total colourings have been studied in only a few articles so far , for example @xcite . for an introduction to circular colouring , as well as a survey of results we refer the reader to @xcite and @xcite . in this work
we study the set @xmath8 , where @xmath9 denotes the circular total chromatic number of the graph @xmath7 .
it is an easy observation to see that @xmath10 . in terms of
other known values of @xmath9 , they are computed precisely for all graphs with maximum degree at most @xmath11 , as well as graphs of order at most @xmath12 , complete graphs , complete bipartite graphs , and some subcubic cases .
in particular it is proved in @xcite that the mbius ladders @xmath13 have circular total chromatic number at most @xmath14 when @xmath15 , and that the prism @xmath16 has circular total chromatic number at most @xmath17 .
we verified by the aid of a computer program that indeed @xmath18 and @xmath19 for @xmath20 . in @xcite an infinite family of graphs with circular total chromatic number at most @xmath17 and a graph with circular total chromatic number at most @xmath21
are given . while it follows from the results of @xcite that the set of circular total chromatic numbers of graphs with maximum degree @xmath11 is precisely the set @xmath22 the only non - integer numbers @xmath23 realized as the circular total chromatic number of some graph are @xmath17 , @xmath14 , @xmath24 and @xmath25 ( see @xcite and the above discussion ) . in particular ,
if @xmath26 , it is not known yet whether the set @xmath27 is infinite . in this work
we prove that every integer @xmath0 is an accumulation point of @xmath28 and hence of @xmath29 .
thus @xmath30 is infinite when @xmath31 .
this work was partly motivated by the following conjecture of xuding zhu for circular edge colouring : @xcite integers @xmath0 are the only accumulation points of the set @xmath32 .
[ conj : zhu ] we conjecture that the same holds for circular total colouring .
for convenience , we allow half - edges in graphs . a half - edge has only one end vertex .
we may join two half - edges to form an edge joining their end vertices .
we use the notion of @xmath33colourings for our upper bounds .
a @xmath33colouring of a graph @xmath7 is any function @xmath34 such that for any pair @xmath35 of adjacent vertices we have @xmath36 .
a @xmath33total colouring of a graph is a @xmath33colouring of its total graph .
the circular total chromatic number of a graph @xmath7 is defined by @xmath37 therefore existence of a @xmath33total colouring for @xmath7 implies @xmath38 .
let @xmath39 be the graph obtained from the complete bipartite graph @xmath40 by deleting one vertex , but leaving its incident edges in as half - edges .
the following lemma is the main ingredient in our construction .
the graph @xmath39 is type @xmath41 .
moreover , in every @xmath42total colouring of @xmath39 , all the half - edges receive the same colour .
[ lem : all0 ] let @xmath43 and @xmath44 be the partite sets of @xmath39 , and let @xmath45 be the half - edge in @xmath39 incident with @xmath46 . throughout this proof @xmath47 and @xmath48 are integers with @xmath49 and @xmath50 .
let @xmath51 $ ] be a latin square of order @xmath52 with entries from the set @xmath53 .
we define a @xmath42total colouring @xmath54 of @xmath39 as follows : @xmath55 , @xmath56 , and @xmath57 .
the fact that @xmath54 is a proper total colouring of @xmath39 follows immediately since @xmath58 is a latin square . for the second part of the lemma , let @xmath54 be a @xmath42total colouring of @xmath39 with the colours @xmath59 .
we may assume that @xmath60 , thus @xmath61 for all @xmath48 .
on the other hand , since each @xmath62 has @xmath52 incident edges , all colours must appear either on @xmath62 itself or on an edge incident with it .
in particular each @xmath62 is incident with an edge coloured @xmath41 .
therefore the set of edges coloured @xmath41 is a matching which saturates @xmath63 , hence it also saturates @xmath64 .
this means that @xmath65 is the only vertex coloured @xmath41 .
subsequently the vertices in @xmath66 must all receive different colours , say @xmath67 for @xmath49 , which leaves only the colour @xmath68 for the vertices @xmath62 .
that is @xmath69 for all @xmath48 .
now the only missing colour at any @xmath46 is the colour @xmath68 , so @xmath70 .
the next lemma is a circular relaxation of lemma [ lem : all0 ] .
c let @xmath31 .
let @xmath71 and @xmath72 be two half - edges of the graph @xmath39 , and let @xmath73 and @xmath74 be the vertices of @xmath39 incident with @xmath71 and @xmath72 respectively .
then for every positive integer @xmath75 , @xmath39 admits an @xmath76total colouring @xmath77 such that @xmath78 , @xmath79 , @xmath80 , and @xmath81 .
[ lem : tweak ] c let @xmath82 and @xmath83 be the partite sets of @xmath39 , and let @xmath45 be the half - edge in @xmath39 incident with @xmath46 .
we may assume @xmath84 and @xmath85 , thus @xmath86 and @xmath87 .
let @xmath51 $ ] be a latin square of order @xmath52 with @xmath88 , @xmath89 , and @xmath90 .
we define a @xmath42total colouring of @xmath39 from @xmath58 as in the proof of lemma [ lem : all0 ] .
we may now multiply all colours by @xmath75 to get an @xmath76total colouring @xmath91 of @xmath39 .
note that in such colouring each occurrence of the colour @xmath68 may be replaced by @xmath92 and the colouring remains proper .
so we may `` tweak '' the colouring @xmath91 by recolouring @xmath93 by the colour @xmath94 .
we then shift all colours by @xmath41 to get the desired colouring @xmath77 .
hence @xmath95 , @xmath96 , @xmath97 , and @xmath98 .
by way of lemma [ lem : all0 ] , we now use the graph @xmath39 to construct type @xmath11 graphs with maximum degree @xmath52 .
let @xmath75 be a positive integer and @xmath99 be the graph obtained from @xmath39 by deleting @xmath100 of its half - edges .
let @xmath101 be all isomorphic to @xmath99 , and let @xmath102 and @xmath103 be the half - edges of @xmath104 .
let @xmath105 be a graph with one vertex @xmath106 and two half - edges @xmath107 and @xmath108 .
we construct the graph @xmath109 from the disjoint union of @xmath110 by joining the half - edges @xmath102 and @xmath111 to form an edge @xmath45 for each @xmath112 .
let @xmath31 .
for all positive integers
@xmath75 , the graph @xmath109 is type @xmath11 .
moreover , @xmath113 .
[ thm : lim ] let @xmath31 and @xmath114 .
let @xmath115 ( @xmath112 ) be as in the definition of @xmath109 .
moreover , let @xmath46 and @xmath116 be the unique end vertices of @xmath102 and @xmath103 respectively . in particular @xmath117 .
suppose @xmath54 is a @xmath42total colouring of @xmath109 .
then by lemma [ lem : all0 ] for each @xmath118 we have @xmath119 .
thus @xmath120 which is a contradiction since @xmath121 and @xmath122 are adjacent at @xmath106 .
therefore @xmath109 is type @xmath11 , which means @xmath123 . to prove @xmath124 , we give an @xmath76total colouring of @xmath109 . by lemma [ lem : tweak ]
, @xmath125 admits an @xmath76colouring @xmath126 with @xmath127 and @xmath128 .
we use @xmath126 to define a ( partial ) @xmath76total colouring of @xmath109 with @xmath129 , @xmath130 , @xmath131 and @xmath132 .
we extend @xmath54 to @xmath104 by shifting the colouring of @xmath125 by @xmath133 .
thus we have @xmath134 for all @xmath112 , and @xmath135 and @xmath136 for all @xmath118 .
we define @xmath137 . to show that @xmath77 is a proper @xmath76total colouring of @xmath109 we only need to show that the new edges @xmath45 do not create any conflict .
note that @xmath138 and @xmath139 are consistent so the edge @xmath45 is valid for @xmath140 . for @xmath121 and @xmath122
we have @xmath132 , @xmath141 which are both consistent with @xmath137 .
finally , @xmath129 and @xmath142 are consistent and since @xmath31 , they are both consistent with @xmath137 .
the following are immediate from theorem [ thm : lim ] . for all @xmath31 , @xmath143 .
[ cor : lim ] for every @xmath31 , the set @xmath30 is infinite .
[ cor : inftly ]
note that @xmath144 is isomorphic to the cycle @xmath145 which has circular total chromatic number @xmath146 as proved in @xcite .
so the upper bound of theorem [ thm : lim ] is tight for @xmath147 . for @xmath26 on the other hand ,
this upper bound is not tight . for a better upper bound
, we need a better tweaking of a @xmath42total colouring of @xmath39 than the rather easy one done in lemma [ lem : tweak ] .
c let @xmath31 .
let @xmath71 and @xmath72 be two half - edges of the graph @xmath39 , and let @xmath73 and @xmath74 be the vertices of @xmath39 incident with @xmath71 and @xmath72 respectively . then for every positive integer @xmath148 , @xmath39 admits a @xmath149total colouring @xmath77 such that @xmath78 , @xmath150 , @xmath151 , and @xmath152 .
[ lem : refine2n ] _ 0 let @xmath82 and @xmath83 be the partite sets of @xmath39 , and let @xmath45 be the half - edge in @xmath39 incident with @xmath46 .
we may assume @xmath153 and @xmath154 , thus @xmath155 and @xmath156 .
let @xmath51 $ ] be the back - circulant latin square of order @xmath52 , namely @xmath157 and @xmath158 .
we start from the @xmath42total colouring of @xmath39 obtained from @xmath58 as in the proof of lemma [ lem : all0 ] , and multiply all colours by @xmath148 to obtain a @xmath149total colouring @xmath77 of @xmath39
. then @xmath159 , @xmath160 , and @xmath161 .
note that the colours used on each vertex and its incident edges ( and half - edges ) are @xmath162 .
thus there is a slack of @xmath41 between @xmath163 and @xmath68 .
so we may shift all the colours @xmath164 by @xmath94 ( @xmath165 ) , or shift all the colours @xmath166 by @xmath41 ( @xmath167 ) and the colouring remains valid at that vertex .
we define @xmath168 from @xmath77 by shifting by @xmath41 at each vertex @xmath46 with @xmath169 , all the colours greater than or equal to @xmath170 . because of our choice of @xmath58 , we have @xmath171 when @xmath172 and @xmath173 otherwise .
effectively , at each vertex @xmath62 all the colours greater than or equal to @xmath174 are shifted by @xmath41 .
therefore @xmath168 is a proper @xmath149total colouring at all vertices . the shifts of colours performed above
enable us now to define @xmath175 and @xmath176 .
we finish by shifting the colours of all vertices , edges and half - edges by @xmath41 .
an example is illustrated in figure [ fig : dist2 ] .
total colouring of @xmath177 found in the proof of lemma [ lem : refine2n ] , before the final shift of all colours .
cthe left column shows the colours @xmath170 , the right column @xmath178 , the top row @xmath179 , and the middle block @xmath180 .
modifications from the original ( scaled ) latin square colouring are shaded . ] for @xmath181 and @xmath114 we have @xmath182 [ thm : improve ]
c let @xmath181 and @xmath114 .
let @xmath183 ( @xmath112 ) be as in the proof of theorem [ thm : lim ] .
we give a @xmath184total colouring of @xmath109 by combining shifts of the colourings of lemma [ lem : refine2n ] . by lemma [ lem : refine2n ] with @xmath185
, we have a @xmath184total colouring of @xmath125 in @xmath109 so that @xmath186 , @xmath187 , @xmath188 and @xmath189 .
for the block @xmath104 where @xmath190 , we use a shift ( by @xmath191 ) of the colouring of @xmath125 .
thus for @xmath118 we have @xmath192 and @xmath193 and @xmath194 .
note that the end vertices of @xmath45 receive colours @xmath193 and @xmath195 which are consistent . at the vertex @xmath106 we have @xmath186 and @xmath196 which are consistent .
moreover , since @xmath197 and @xmath198 , we may define @xmath199 .
this does not conflict any of the adjacent colours since @xmath181 . combining @xmath200 shifts of an @xmath201total colouring of @xmath202 with @xmath203 , @xmath204 ( obtained from lemma [ lem : refine2n ] ) and an @xmath201total colouring of @xmath202 with @xmath203 , @xmath205 , @xmath206 , @xmath207 ( obtained by tweaking the proof of lemma [ lem : tweak ] ) , one can prove @xmath208 .
we omit details of such proof here . by computer aid
, we verified that the upper bound of theorem [ thm : improve ] is tight for @xmath209 where @xmath210 and @xmath211 where @xmath212 .
we also verified by computer aid that @xmath213 and @xmath214 for @xmath215 .
these results suggest that the upper bounds of this section could indeed be the actual circular total chromatic numbers of the graphs @xmath109 .
similar to circular edge colouring , the circular total chromatic numbers of graphs seem to be sparse . by analogy to conjecture [ conj : zhu ] and based on our computational experiments , we conjecture the following .
it is worth mentioning here that the graph @xmath109 is bipartite when @xmath75 is odd .
hence corollary [ cor : lim ] and corollary [ cor : inftly ] remain valid for bipartite graphs .
a natural question to be asked here is whether every member of @xmath29 can be realized as the circular total chromatic number of some bipartite graph . | for every integer @xmath0 and every @xmath1 we construct a graph with maximum degree @xmath2 whose circular total chromatic number is in the interval @xmath3 .
this proves that ( i ) every integer @xmath0 is an accumulation point of the set of circular total chromatic numbers of graphs , and ( ii ) for every @xmath4 , the set of circular total chromatic numbers of graphs with maximum degree @xmath5 is infinite .
all these results hold for the set of circular total chromatic numbers of bipartite graphs as well . |
go solution was prepared using a modified hummers method with additional pre - oxidation of the graphite flakes .
@xmath98 of the graphite powder were pre - oxidized using sulfuric acid ( @xmath99 ; @xmath100@xmath40 @xmath101 ) , phosphorus pentoxide ( @xmath102 ; @xmath103 ) , and potassium persulfate ( @xmath102 ; @xmath104 ) in a water bath at @xmath105 for @xmath106 .
the mixture was then cooled to room temperature and diluted with @xmath107 of deionized water ( di ) and vacuum filtered through a poly(tetrafluoroethylene ) membrane ( pore size @xmath108 ) .
after pre - oxidation , the modified hummers method was carried out by suspending the pre - oxidized material in @xmath101 ( @xmath109 ; @xmath100@xmath40 ) in an ice bath for @xmath110 .
potassium permanganate was slowly added ( @xmath111 ; @xmath112 ) and the solution was heated to @xmath113 for @xmath114 .
the process was completed by adding @xmath115 of di water and heating the mixture to @xmath116 for an additional @xmath114 .
the reaction was quenched with hydrogen peroxide ( @xmath99 ; @xmath117 ) and di water ( @xmath118 ) . in order to quench the unreacted reagent and
clean the solution , the product was then filtered through a @xmath119 testing sieve , then through glass fiber and centrifuged for @xmath120 at @xmath121 ( sorvall rc-5c plus ) .
the supernatant was discared whereas the precipitate was washed with hydrochloric acid ( @xmath122 , @xmath123@xmath40 @xmath124 ) .
the sieving , centrifugation , and washing process was then repeated using di ( two times ) and ethanol ( @xmath125 , two times ) .
the final solution was then dispersed in @xmath126 of @xmath125 at a concentration of @xmath127 .
the overall yield of the process was circa @xmath128@xmath40 .
figure s[s1 ] represents a sem image of a drop casted go solution on a silicon wafer .
as one can see , more than @xmath129 @xmath40 of the go flakes is monolayer with an average area of @xmath130 .
figure s[s2 ] represents sem images obtained after depositing @xmath131 @xmath132 layer onto the membranes via an ems300r sputter coater .
figure s[s2]a shows the bare pvdf membrane characterized by pores size of @xmath133 .
figure s[s2]b represents the pvdf membrane after the go deposition via vacuum filtration .
figure s[s2]c shows a cross section of the membrane , which highlights the extreme thinness of the go layer ( circa @xmath134 ) .
it is also possible to observe that the go layer conserves the morphology of the underneath pvdf membrane .
as explained in the main text , the xrd spectra was obtained by integrating a 2d diffraction pattern image ( fig .
s[s3]a ) . the xrd spectra ( fig .
s[s3]b ) displays a distinct peak centered at @xmath135 . by using bragg s law
, this value can be converted to @xmath136 , which represents the spacing between the go flakes .
the permeability tests were carried out with a custom - made dead - end filtration system ( figure s[s4 ] ) .
the filtration system includes a @xmath137 reservoir ( e.g. pressure pot ) .
the pressure was regulated with a pressure gauge from ingersoll and the go membrane was placed inside a stainless steel edm millipore filter holder . finally .
s[s5 ] displays the normalized flowrate versus applied pressure for the go membrane . as mentioned in the main text ,
the permeability is independent from the applied pressure and this is testified by the linear fitting .
the possibility of slip flow was contemplated by navier himself as early as in 1823 , by postulating a direct proportionality between the slip velocity and the gradient of the velocity at the wall .
for a one - dimensional flow across parallel plates at a distance @xmath79 , the navier slip - boundary adds a slip term @xmath139 , on top of the parabolic poiseuille bulk profile @xmath140 .
one may therefore argue that the notion of slip - flow is perfectly compatible with continuum hydrodynamics .
this is certainly true for the case of macroscopic flows , in which the slip term @xmath141 .
however , for nanoflows , where @xmath79 becomes comparable with @xmath142 , this is no longer true , because the slip term becomes dominant over the bulk one .
moreover , under such conditions , it is not obvious that the bulk component still obeys a parabolic flow profile , because @xmath143 means that non- equilibrium is as strong as equilibrium , thus contradicting the low - knudsen assumption , @xmath141 , which lies at the basis of the continuum picture .
a.m. , g.f . and
were partially supported by the integrated mesoscale architectures for sustainable catalysis ( imasc ) , an energy frontier research center funded by the us department of energy , office of science , basic energy sciences under award no .
de - sc0012573 .
c.a.a . would like to thank computational and fluid dynamics class ( ap274 ) at harvard , from where this work started .
this work made use of the center for nanoscale systems at harvard university , a member of the national nanotechnology infrastructure network , supported ( in part ) by the national science foundation under nsf award number ecs-0335765 .
37ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.101.226101 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1021/acsami.5b12112 [ * * , ( ) ] , , @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) | the understanding of water transport in graphene oxide ( go ) membranes stands out as a major theoretical problem in graphene research .
notwithstanding the intense efforts devoted to the subject in the recent years , a consolidated picture of water transport in go membranes is yet to emerge . by performing mesoscale simulations of water transport in ultrathin go membranes ,
we show that even small amounts of oxygen functionalities can lead to a dramatic drop of the go permeability , in line with experimental findings .
the coexistence of bulk viscous dissipation and spatially extended molecular friction results in a major decrease of both slip and bulk flow , thereby suppressing the fast water transport regime observed in pristine graphene nanochannels . inspection of the flow structure reveals an inverted curvature in the near - wall region , which connects smoothly with a parabolic profile in the bulk region .
such inverted curvature is a distinctive signature of the coexistence between single - particle langevin friction and collective hydrodynamics .
the present mesoscopic model with spatially extended friction may offer a computationally efficient tool for future simulations of water transport in nanomaterials .
water transport through graphene - derived membranes has recently captured much attention due to its promising potential for many technological applications , such as water filtration , separation processes , and heterogeneous catalysis @xcite .
a number of simulations have shown that fast water permeation through carbon materials , such as carbon nanotubes ( cnt @xcite ) and pristine graphene membranes @xcite is due to the slip flow at the water - carbon interface .
fast water transport ( fwt ) with permeabilities higher than @xmath0 @xcite has also been reported for graphene oxide laminate ( gol ) .
+ however , in contrast to pristine graphene and cnt , a clear consensus on the gol fast water transport mechanism has yet to emerge @xcite . to date ,
departures from hydrodynamic ( i.e. , hagen - poiseuille ) behavior are typically attributed to the low friction experienced by water in atomistically smooth graphene nanochannels @xcite or to the presence of structural defects in gol , leading to shorter water paths @xcite .
the latter hypothesis is further corroborated by the breakdown of fwt from molecular interactions of water with basal plane hydroxide and epoxide groups , which hinders the motion of water molecules .
this phenomenon was recently reported in non - equilibrium molecular dynamics simulations of flows inside graphene oxide ( go ) nanochannels , @xcite . under such conditions ,
slip - corrected continuum hydrodynamics is expected to provide a correct description of the aforementioned transport mechanism .
however , at the hydrodynamic level the details of the slip corrections must be imposed a priori because the physics of slip flow is governed by non - equilibrium phenomena .
such details can certainly be accounted for by a fully atomistic approach , but at the price of prohibitive computational costs .
all of the above lead to an ideal scenario for _ mesoscale _ modeling techniques , which may offer a valuable compromise between physical fidelity and computational viability , @xcite .
+ in this work , we explore water permeation in gol via mesoscale simulations .
furthermore , we provide a direct calculation of the membrane s permeability , which is drastically lower than the one observed in pristine graphene characterized by fwt .
our simulations exhibit satisfactory agreement with experimental gol permeability and also provide values of the slip length on par with recent non - equilibrium molecular dynamics simulations ( nemd ) @xcite .
more precisely , we demonstrate that a suitably simplified ( lattice ) kinetic model proves capable of predicting the breakdown of fwt , namely the macroscopic permeability of the go membrane , as well as the slip phenomena , at a very affordable computational cost . for instance , a baseline simulation takes only a few cpu hours on a high - end pc .
* method .
* go was synthesized by modified hummers method @xcite and dispersed in ethanol at a concentration of @xmath1 .
additional details can be found in the supporting information .
the solution was then bath sonicated for 5 minute in a branson 2510 ultrasonic cleaner .
@xmath2 of the sonicated go solution was deposited on a porous polyvinylidene difluoride ( pvdf ) membrane ( pore size = @xmath3 ) via vacuum filtration .
the membranes were then characterized by using a zeiss ultra field emission scanning electron microscope ( sem ) with an in - lens detector ( see figure s1 and s2 in the supporting information )
. scanning electron microscopy and mass analysis suggest the formation of a circa @xmath4 thick go layer on top of the pvdf membrane .
the crystallographic structure of the membranes was analyzed with a bruker d8 equipped with a two - dimensional vantec-500 detector .
the spectra were obtained by the integration of the 2d diffraction pattern image ( see figure s3 in the supporting information ) via eva software . depending on the sample , the integration time was between 600 - 1200 seconds .
permeability results were carried out with a custom - made dead - end filtration system operating at a maximum pressure of @xmath5 . in particular , a @xmath6 in diameter go membrane was cut and placed in a stainless steel edm millipore filter holder .
the pressure was monitored with an ingersoll pressure gauge regulator ( see figure s4 in the supporting information ) .
the numerical simulations are based on the lattice boltzmann ( lb ) method , augmented with a novel langevin - like frictional force accounting for the go - water interactions .
since the lb method is largely documented in the literature @xcite , in the following we discuss only its basic features .
the lattice boltzmann equation reads as follows : @xmath7 where @xmath8 is a set of discrete probability distribution functions ( pdfs ) representing the probability of finding a molecule at position @xmath9 and time @xmath10 with a lattice - constrained velocity @xmath11 , where the index @xmath12 runs over the nine directions of the lattice @xcite .
@xmath13 is the set of discrete local maxwellian equilibria ( i.e. , truncated low - mach number expansion of the maxwell - boltzmann distribution ) @xcite and @xmath14 is the speed of sound of the model @xcite . in the above equation ,
the left hand side is the lattice transcription of molecular free - flight along the lattice directions , while the right hand side describes collisional relaxation towards local equilibrium , described by a low - mach number expansion of the maxwell - boltzmann distribution .
the relaxation parameter @xmath15 controls the kinematic viscosity of the lattice fluid through the relation @xmath16 in lattice units @xmath17 .
the last term , @xmath18 , is the correction due to the friction exerted by the hydroxyl and epoxy ( see fig . [
hydroxyl]b ) on the water molecules and can be expressed as : @xmath19 , in which @xmath20 is the set of weights for the chosen lattice @xcite , and the frictional force is taken in the following langevin form @xcite : @xmath21 where @xmath22 and @xmath23 are the fluid density and velocity respectively and the heterogeneous friction coefficient reads as follows : @xmath24 where w is a representative size of the protruding functional groups and @xmath25 is a characteristic water - hydroxyl collision frequency , taken equal to @xmath26 ( in lattice units ) in all the simulations , corresponding to a collision rate of about @xmath27 @xcite .
the wall function @xmath28 takes the value @xmath29 for @xmath30 and @xmath31 elsewhere .
likewise , @xmath32 for @xmath33 and @xmath31 elsewhere .
the reference case is @xmath34 ( truncated langevin throughout the text ) . in this case , if @xmath35 , the langevin force drops discontinuously to zero in the central region of the channel , w@xmath36w .
to regularise this discontinuity , we also consider the case @xmath37 ( non - truncated langevin ) .
the physical idea behind the langevin - like frictional force is to account for the complex water - go molecular interactions at a coarse - grained level , whereby all atomistic details are channeled into the parameters @xmath25 and w. whether the contact angle and the slip length can be treated as independent variables , still is an open question in the current literature ( see @xcite ) . in line with the mesoscale nature of our model , we assume sufficient universality to support a direct correlation between the water - graphene contact angle and the slip length .
this said , from the nemd results ( fig .
4d in @xcite ) , we read off a slip length between @xmath38 with a @xmath39 @xmath40 to 5@xmath40 of hydroxyl groups , respectively . according to the authors , these values correspond to a contact angle @xmath41 , with 20@xmath40 hydroxyl groups and a @xmath42 slip length in pristine graphene . based on these data
, it is reasonable to assume that the slip length should be of the order of the molecular size of the hydroxyl groups ( @xmath43 ) , which is precisely the assumption made in our model .
this phenomenological approach allows us to import knowledge at the microscale within a mesoscale computational harness .
in particular , the model permits us to reach space - time scales of experimental relevance without being tied - down to the continuum assumptions the lattice units are @xmath44 , yielding a time step @xmath45 .
note that sub - molecular spatial resolutions are typical of lb simulations of nanoflows @xcite .
the time - step , however is about an order of magnitude larger then the time steps typically used in md simulations of flows through go interlayers @xcite .
it is also worth mentioning that the cpu time needed to update a single molecule is significantly larger than the one required to advance a single lb cell , because the cell contains less neighbours and , more importantly , such neighbours are fixed in space , hence there is no need to recompute the interaction list at every time - step .
. moreover , lb requires no statistical averaging , since it is based on a pre - averaged probability distribution function .
in addition , lb is often more efficient than computational fluid dynamics because i ) pressure is available locally in space and time , with no need of solving a demanding poisson problem , ii ) transport is exact , since free - streaming proceeds along fixed molecular velocities instead of space - time dependent material streamlines , iii ) diffusion is emergent , hence it does not require second order spatial derivatives , thus facilitating they formulation of boundary conditions in confined flows @xcite .
* results . * assuming the gol structure to be symmetric and periodic @xcite ( see red area in fig .
[ hydroxyl]a ) , we consider only two nanochannels out of the full device . a similar geometry set - up has been recently employed to investigate water permeation through graphene - based membranes by means of md simulations @xcite .
these two channels are connected via two openings of half the width of the inlet / outlet pores ( see fig .
[ hydroxyl]a ) .
the boundary conditions at the left - right and top - bottom surfaces are periodic , to simulate the proximity of two adjacent go layers . at solid walls ,
the molecules experience the langevin friction previously described . and channel spacing @xmath46 .
the permeate increases linearly with time , thus denoting a hydrodynamic ( darcy - like ) behaviour of the membrane .
see also the flow rate vs the pressure gradient plot reported in figure s5 of the supplementary information , reporting a linear behaviour typical of the hydrodynamic regime .
dashed line indicates the compaction time of the membrane . ]
we have run several simulations with different values of w. the grid resolution is @xmath47 , corresponding to a flake length of @xmath48 , which we take as a representative experimental value , @xcite .
the mesoscale model was tested against experimental measurements of permeability .
the permeability of the membrane is defined as @xmath49 , where @xmath50 is the dynamic viscosity , @xmath51 the pressure gradient across the membrane and @xmath52 the superficial velocity defined as the ratio between the membrane discharge per unit area .
it is worth recalling that in the darcy regime , the permeability @xmath53 is pressure independent ( see figure s5 in the supporting information ) . in order to test our numerical results against experimental data , we inferred the experimental values of flow rates ( mass flow per unit time ) from the permeate vs time plot , reported in fig .
[ permeate ] , for different values of the applied pressure . from these data , we compute a dimensionless permeability @xmath54 , where @xmath55 represents the spacing between two go layers .
x - ray diffraction measurements give @xmath56 . the experimental value of the dimensionless permeability @xmath57 for a @xmath58 thick membrane of diameter @xmath59 is @xmath60 , corresponding to a permeability of @xmath61
this value of permeability is in line with data previously reported for ultrathin ( @xmath62 ) go membranes ( see @xcite ) and represents a promising result for nanofiltration applications .
it is important to underline that higher values of permeability have been achieved for go membranes .
however , these values can be connected to the presence of defects or larger go nanochannels due to the chemical modification of the go membranes @xcite .
higher permeability can also be achieved by intercalating the go membranes with high aspect ratio nanoarchitectures , such as cnt @xcite .
fig .
[ kappa ] illustrates two comparisons , for both the truncated and non - truncated langevin frictions : i ) between the permeability obtained by mesoscopic simulations and experiments ; ii ) between the slip length obtained by nemd @xcite and by this study .
the case w@xmath63 corresponds to a simulation without the langevin friction , i.e. pure _ free - slip _ hydrodynamics . by free - slip we mean boundary conditions which leave the flow momentum tangential to the wall unchanged
this is representative of the fwt regime observed in pristine graphene experiments .
note that the flow still reaches a steady - state solution on account of the localized dissipation experienced at the sharp @xmath64 turns between two subsequent layers , visible in fig .
[ hydroxyl]c . the main outcome from fig .
[ kappa ] is a dramatic drop in permeability of two orders of magnitude , already at w@xmath65 , i.e. w@xmath66 .
such a dramatic drop in permeability is consistent with experimental observations that report a suppression of fwt regimes in the presence of hydroxyl groups .
further increments of w lead to a sizable reduction in permeability by circa one order of magnitude , from w@xmath67 to w@xmath68 .
it is worth noting that , the employed resolution ( @xmath69 ) poses a constraint on the lower bound of w. we note that the truncated ( @xmath70 , @xmath34 ) and non - truncated ( @xmath71 , @xmath37 ) scenarios yield nearly the same picture with only minor quantitative variations .
moreover , fig .
[ kappa ] shows that for both the truncated and non - truncated scenario , the slip length remains between @xmath72 and @xmath73 , in agreement with the values provided by nemd @xcite .
it is worth noting that the best match of the experimental ( permeability ) and nemd ( slip lengths ) results with the simulations is obtained when w is between @xmath74 and @xmath75 ( i.e. w@xmath76 ) , which agrees with the physical dimension of the oxygen functionalities in the go nanochannels @xcite .
this further corroborates the validity of the model and its potential to capture the physical phenomena in 2d nanostructured inspired materials within an efficient computational framework .
* inside the flow structure . *
next , we proceed to inspect the internal structure of the flow . in particular , we focus on the occurrence of slip flow in the presence of hydroxyl , as recently observed in non - equilibrium molecular dynamics simulations @xcite . as previously discussed , slip flow is a typical manifestation of individual non - hydrodynamic behavior driven by non - equilibrium effects near the wall . to glean quantitative insights into such non - equilibrium phenomena is informative for inspecting the one - dimensional cross - flow profiles @xmath77 for different values of w versus the case of free - slip boundary conditions .
fig .
[ flowprofiles]a illustrates that the flow profiles display a poiseuille - like parabolic trend in the bulk region , smoothly turning into a flat profile near the wall , with a positive curvature and a small non - zero slip velocity .
the comparison between the pristine graphene profile with the other profiles highlights the major drop of mass flow due to langevin friction .
the increase of w leads to the suppression and flattening of the water profiles . from the velocity profiles , we compute the slip length according to @xmath78 . as previously discussed for fig .
[ kappa ] , we confirm the presence of a residual slip length , which is small in absolute physical units , but fairly sizable in units of the channel width @xmath79 , namely @xmath80 .
furthermore , the effect of the different cut - off lengths employed in the truncated @xmath81 and non - truncated @xmath82 langevin is minor , leading to slight differences in both the membrane permeability and slip lengths . to further test the robustness of this approach in fig .
[ flowprofiles](b ) we report a comparison between the velocity profiles obtained by the langevin - lb and the md approach @xcite for a @xmath83 wide go channel , showing an outstanding agreement between the two models .
we conclude this letter with an ex - post analysis of our mesoscale model in light of the numerical results discussed above .
first , we note that although both friction and viscous dissipation withstand the driving action of the pressure gradient , they operate according to very different and competing mechanisms .
friction drives the fluid towards the following local flow configuration : @xmath84 this results in a slip - flow @xmath85 ( independent of w ) at the wall and the corresponding slip length is @xmath86w .
viscous dissipation , on the other hand , drives the fluid towards a parabolic hagen - poiseuille profile , @xmath87.\ ] ] where we have set @xmath88 .
the above hydrodynamic solution is compatible with either slip or non - slip flow conditions , depending on the value of the constant at the right hand side . in slip - hydrodynamics @xmath89 , so that @xmath90 recovers standard no - slip hydrodynamics .
this value can only be prescribed a - priori , exposing the weakness of the hydrodynamic approach previously mentioned .
the two profiles , @xmath91 and @xmath92 can not coexist other than in an intermediate compromising form , resulting from the smoothing of the langevin profile due to viscous dissipation . to gain a deeper understanding , in fig .
[ figforce ] we report the friction and viscous forces separately , for the case w@xmath93 .
as one can see , the bulk flow is dominated by viscous dissipation ( friction is zero in the bulk because @xmath94 ) , while in the vicinity of the wall , friction takes over .
however , the two contributions become comparable but opposite in sign due to the positive curvature of the flow profile .
we have also checked through separate simulations that the sign inversion of hydrodynamic dissipation holds true at higher resolution ( see fig . [ flowprofiles](b ) ) .
the inversion of the dissipative force , namely @xmath95 at the wall is a genuine effect of the spatially distributed decaying langevin friction . indeed , since the free - slip boundary condition forces @xmath96 , the near - wall dissipative force is dominated by the langevin friction , which is incompatible with @xmath97 . since langevin friction decays exponentially away from the wall , the near - wall flow profile must grow exponentially according to eq.[ul ] , thus exhibiting the observed inverted ( positive ) curvature .
because of the exponential decay of the extended friction , the bulk region is still dominated by viscous dissipation , as reflected by the bulk parabolic profile clearly visible in fig.[figforce ] .
the onset of the inverted curvature is a distinctive signature of the coexistence between single particle langevin friction ( slip - flow ) and collective hydrodynamics ( bulk flow ) . differently from continuum methods , which impose the slip length through the boundary conditions , in our model it naturally emerges from the self - consistent competition between extended friction and hydrodynamic dissipation .
this extra - freedom is key to recover the inverted curvature profile .
* conclusions . * in summary ,
our numerical simulations portray the following picture .
even a small amount of spatially extended langevin friction , w@xmath66 , leads to a dramatic drop in the mass flow compared to free - slip hydrodynamics .
such friction still supports a small residual slip - flow at the wall , with a slip length of the order of the friction length w. however , this flow is largely negligible compared to the free - slip in the absence of langevin friction .
the net result is a dramatic loss of permeability due to the presence of the functional groups , hence the inhibition of the fwt regime observed in pristine graphene membranes .
viscous effects dominate the bulk flow and contribute to the smoothing of the sharp features of the flow due to the presence of the hydroxyl and epoxy .
inspection of the flow structure reveals an inverted curvature in the near - wall region , which connects smoothly with a parabolic profile in the bulk region such inverted curvature is a distinctive signature of the coexistence between single particle langevin friction and collective hydrodynamics . |
the x - ray transient v4641 sgr ( sax j1819.32525 ) is a unique black - hole binary , which has attracted considerable interest since its major outburst in september 1999 , where it peaked at @xmath5 ( stubbings 1999 ) . even in quiescence ,
the star is fairly bright at @xmath6 mag .
it has an orbital period of 2.817 days , and is located at a distance of @xmath7 kpc toward the galactic bulge ( orosz et al .
the total mass of the system lies between 1420 m@xmath8 with a black hole primary of @xmath9 m@xmath8 ( orosz et al .
given its moderate inclination ( 6070@xmath10 , orosz et al .
2001 ) , the system does not show eclipses and the quiescence light curve is dominated by ellipsoidal variability .
the secondary was classified as a b9iii star , being one of the most luminous secondary stars known in dynamically established black - hole binaries ( orosz et al .
the large eruption in 1999 was associated with superluminal jet ejection and was followed by a series of smaller outbursts in the last few years .
these involved violent flashes and rapid variability , making this relatively bright system a key target for understanding the processes associated with black hole binaries .
the activity of v4641 sgr has been variable since the 1999 september outburst . as a faint x - ray source
it was discovered independently by rxte ( markwardt et al . 1999 ) and bepposax ( int zand et al . 1999 ) in 1999 february .
the 1999 september event , peaking at 12.2 crab in the 212 kev band , was extremely rapid in its rise and decay across radio , optical and x - ray wavelengths . in x - rays ,
a brief but dramatic eruption occured on sept .
15 - 16 , 1999 , showing remarkably rapid fluctuations ( wijnands & van der klis 2000 ) .
the optical decay lasted about two days ( e.g. chaty et al .
2003 ) , while the radio transient was detected for a further three weeks .
the measurable proper motion of the resolved radio structures indicated superluminal jet ejection ( hjellming et al . 2000 ) .
subsequent quiescence observations led orosz et al .
( 2001 ) to measuring the most accurate system parameters so far , which revealed the massive stellar black hole primary and the nature of the early - type secondary star .
the second major outburst occurred in 2002 may , when unprecedented rapid optical fluctuations were detected ( uemura et al .
the short time - scale of the large - amplitude variations indicated that the inner region of the accretion disk made a significant contribution to the optical flux , implying strong non - thermal emission in the optical ( uemura et al .
2002 , 2004b ) .
another outburst in 2003 august was very similar to the one in the previous year , leading uemura et al .
( 2004a ) to the conclusion that the two eruptions had the same nature .
they also suggested the possibility of recurrent outbursts on a time - scale of 12 years .
this estimate of the time - scale was supported by r.stubbingss visual report of a fourth major outburst on 2004 july 4.368 ut , where the star peaked at around @xmath11 mag ( uemura et al . 2005 ) . coincidentally
, the star was detected with the rxte at 8.2 mcrab flux ( 2 - 10 kev ) on july 3.496 , roughly eight times fainter than during the outburst in 2003 ( swank 2004 ) and three orders of magnitude dimmer than in the large 1999 september event .
vla observations clearly detected the outburst in radio , as a 12.5 mjy source at 1.425 ghz , with a steep spectrum ( rupen et al .
a day later , v4641 sgr was already close to its quiescence level ( revnivtsev et al .
2004 ) , although very rapid , post - outburst flares of 1 mag were seen as late as 17 days after the main outburst ( bikmaev et al .
2004 , uemura et al .
most recently , a new active phase occured in late june , 2005 , when v4641 sgr was detected in a very similar outburst like those in 2002 , 2003 and 2004 ( swank et al .
2005 , khamitov et al .
2005 , buxton et al . 2005 ) .
we started spectroscopic monitoring of the system on 2004 july 4.48 , roughly two hours after the 2004 outburst was reported via electronic mailing lists , with a few optical and near - infrared spectra ( an early report was given by kiss & mszros 2004 ) . in all follow - up observations between 2004
september and 2005 march were taken in the quiescence of v4641 sgr and they allowed us to measure an improved set of spectroscopic elements .
some of the important questions we wish to address in this paper are as follows .
what is the driving mechanism behind these recurrent outbursts ?
what sort of processes accompany an eruption ?
how much are the 20022004 outbursts different from the major one in 1999
september ? which system parameters can be better constrained using our dataset ?
we describe the observations in sect . 2 ; sect .
3 discusses the outburst spectra , while quiescence data are presented in sect .
discussion of the above questions can be found in sect . 5 and the paper concludes in sect .
[ cols="<,<,^,^,^,^,^,^ " , ] the observations were carried out with the 2.3 m telescope at the siding spring observatory , australia .
four sets of observations were obtained : 2004 july 4 , during the outburst ; four nights in late 2004 september , nine nights in 2004
october / november and six nights in 2005 march , all in quiescence ( see table [ log ] for details ) .
all spectra were taken with the double beam spectrograph using 1200 mm@xmath2 gratings in both arms ( except in 2005 march , when a ccd failure prevented the use of the blue arm ) .
the exposure time ranged between 240900 s for the outburst spectra and was 1200 s in quiescence .
the dispersion was 0.55 px@xmath2 , leading to a nominal resolution of about 1 .
in addition to the v4641 sgr spectra , we regularly observed a telluric standard ( hd 177724 ) and several radial velocity standard stars ( hd 187691 , @xmath12 vir , hr 3383 ) .
the latter were used to measure cross - correlation radial velocities in quiescence .
all spectra were reduced with standard tasks in iraf .
reduction consisted of bias and flat field corrections , aperture extraction , wavelength calibration and continuum normalization .
we did not attempt flux calibration because the conditions were generally non - photometric .
we checked the consistency of wavelength calibrations via the constant positions of strong telluric features , which proved the stability of the system .
radial velocities were determined with the task _ fxcor _ , including barycentric corrections .
different velocity standards have shown that our absolute velocity frame was stable to within @xmath1323 km s@xmath2 .
about two hours after the first visual report arrived , we took three spectra in both the blue and red arms ( between july 4.48 - 4.50 ut ) .
the latter was observing in the near - infrared , showing spectra dominated by the paschen series .
five hours later , we went back to the star to take another near - infrared spectrum , after which we switched to the h@xmath0 region .
thereafter we took spectra in both arms continuously until the object was too low over the horizon .
we obtained a total of 24 spectra , which showed remarkable changes with time .
orbital phases were assigned using the following ephemeris for the shallower minimum of the ellipsoidal variability ( uemura et al . , in preparation ; see also sect . 4 for the spectroscopic ephemeris ) :
@xmath14 using this ephemeris , the orbital phases covered by our data ranged between 0.89 and 0.99 , where 1.00 corresponds to the secondary passing in front of the black hole .
we show a representative collection of the three spectral regions observed during outburst in figs [ tp1][tp2 ] and fig .
lines of hydrogen and helium dominate the spectrum with weaker detections of fe ii , mg ii and si ii .
rapid variations in line profiles between spectra indicate a high and unpredictable activity , whereas a large range of gas velocities is reflected in the broad emission profiles . here
we identify the strong blend of the he i @xmath155015 and @xmath155048 lines and four fe ii lines , of which the fe ii @xmath155169 is the strongest ( fig . [ tp1 ] ) , while fe ii @xmath155235 and fe ii @xmath155276 are marginally detected .
line profiles of he i @xmath155015 exhibited a rapid change between the early and the late phase , separated by more than five hours . during the early phase , it contained a red - shifted peak , a deep absorption and a blue - shifted peak .
the blue peak apparently has a narrow peak with a broad shoulder . during the late phase ,
the line profile becomes a complex emission line with a strong red peak , a blue component and a broad blue wing .
the full width at zero intensity ( fwzi ) reached about 5000 km s@xmath2 , although it is difficult to measure due to the close blend .
of the four fe ii lines , the strongest fe ii @xmath155169 line showed a clear p cyg - like profile in the early phase and possibly in the late phase .
similar features are also visible in the fe ii @xmath155317 line . the absorption components were blue - shifted by 500600 km s@xmath2 , relative to the emission peak .
region at july 4.727 ut ( black / solid line ) and july 4.765 ut ( red / dashed line).,width=302 ] line.,width=302 ] this region was dominated by a very broad and strong h@xmath0 line , the he i @xmath156678 line and weak mg ii / si ii emission features .
the fwzi of the h@xmath0 line exceeded 5000 km s@xmath2 , with an equivalent width ( ew ) of about 130140 .
the asymmetric line profile of the h@xmath0 line can be explained by the presence of a few ( at least three ) distinct components , namely a strong red - shifted component , a weaker blue - shifted component and a broad wing .
the he line had approximately the same structure .
both profiles suggest the presence of a high - velocity outflow component , which is strikingly similar to what was reported in the 1999 september outburst ( orosz et al .
2001 , chaty et al .
2003 ) .
line ( red crosses with dotted line ) and observed apparent magnitudes ( solid line ) .
the typical error of the equivalent width is a few . _
middle panel : _ time variations of the fwhms of the three gaussian components , expressed in km s@xmath2 .
blue and red crosses show the blue and red components , while open squares refer to the broad component . _
bottom panel : _ time variations of the line centres of the gaussian components .
symbols as in the middle panel . ,
width=321 ] the roughly one hour of continuous observations in the h@xmath0 region revealed further interesting features .
firstly , there was a significant change in the line profile over the one hour interval .
secondly , from fig .
[ tp2 ] , it is apparent that the red peak moved further redward , while the blue component moved significantly blueward .
the width of the red component also increased with time , while very similar changes occurred in the he i @xmath156678 line . to quantify these changes , we fitted multiple gaussian components to all 10 h@xmath0 profiles .
a sample fit is shown in fig .
it turned out that the red peak moved by @xmath1680 km s@xmath2 towards red , while the blue peak moved almost twice as much , @xmath16150 km s@xmath2 towards blue .
their widths also changed from spectrum to spectrum , indicating significant changes in line profiles . in parallel with these h@xmath0 spectra , one of us ( m.u . ) made simultaneous ccd photometric observations in chile , using a 30 cm telescope at universidad de concepcion .
these observations covered a 10-minute flare with 0.2 mag amplitude at jd=2,453,191.233 , which just coincided with taking two spectra in australia .
the one - hour interval of simultaneous photometry and spectroscopy allowed us to investigate unprecedented detail into the rapid variability of the system ( the full photometric behaviour of the 2004 eruption is discussed in uemura et al .
2005 ) .
we correlate the equivalent width of the h@xmath0 line and parameters of the fitted gaussian components with the light variations in fig .
[ flare ] .
the top panel shows that there was an inverse correlation between the system s apparent magnitude and the h@xmath0 equivalent width .
this indicates that the flare - up was largely a flare of continuum .
the middle panel in fig .
[ flare ] suggests , however , that line profile changes also accompany the flaring activity : of the three components , the blue and the broad ones narrowed significantly in the flare maximum .
the formal error of the widths of the gaussian fits is about 2 - 3% , which was definitely exceeded by the observed @xmath17% changes .
it is very interesting that the width of the red component was very stable apart from a gradual broadening with time , indicating that it was formed in a completely different part of the system that did not take part in the flaring activity . finally , the bottom panel in fig .
[ flare ] shows the variations of the doppler - shifts of the components with time .
the individual errors are between 1050 km s@xmath2 , so the observed changes are real .
interestingly , they do not show any visible correlation with the flaring activity , suggesting the lack of strong asymmetries in line profile variations . in the near - infrared spectra
we identify higher order lines of the paschen series , all in emission , the he ii @xmath158237 line with a p cygni profile and possibly some emission of the o i @xmath158446 line . due to the broadness of the features , there might be additional contamination from other species .
for example , the p16 , 15 and 13 lines could be contaminated by calcium emmissions at 8498 , 8542 and 8652 , respectively.we show a zoomed plot of the paschen lines in fig .
[ tp3 ] , where the dashed - line spectrum was taken 5 hours later .
the paschen lines changed significantly over the 5 hours between july 4.487 and 4.718 ut .
these changes were similar both to those of the he i lines in the blue region and the he i @xmath156678 line in the h@xmath0 region , showing stronger red - shifted and blue - shifted emission components .
the fwzi increased from 10001500 km s@xmath2 up to 3000 km s@xmath2 , which is also similar to that of the he i @xmath156678 line .
finally , whereas the earlier spectrum showed comparable line strengths for the paschen lines , this changed dramatically later , with the lower members of the series ( p11 and p12 ) becoming much stronger than the other lines .
this is indicative of switching from optically thick to optically thin emission , where the latter is characterized by a regular flux increase toward the lower member of the series ( see storey & hummer 1995 for theory and clark et al .
1998 for an application in a be / x - ray binary ) .
we return to the implications of these spectra in sect . 5 .
we continued regular observations of v4641 sgr after it returned to quiescence .
our monitoring has resulted in further 46 spectra , taken between 2004 sept . 25 and 2005
none of the spectra showed any sign of variability or spectral line distortion that might have occurred due to some accretion activity ( goranskij et al .
2003 ) , with the only detectable change due to the orbital motion of the second component .
given that the one published radial velocity curve ( orosz et al .
2001 ) had quite large uncertainties ( see fig . 2 in orosz et al .
2001 ) , we decided to re - determine spectroscopic elements of the system . a sample quiescence spectrum is shown in fig .
[ quiesc ] .
it is dominated by the strong h@xmath0 absorption line , which is characteristic for a late b / early a type star .
the strong na i d doublet is of interstellar origin as its radial velocity does not change with the orbital phase .
apart from these , only weaker features of mg ii and si ii are easily identifiable , which agrees very well with the quiescence observations of orosz et al .
( 2001 ) . as a comparison
, we also plot a spectrum of hr 3383 , an a1v star .
the close similarity of the spectra supports the spectral type determination by orosz et al .
( 2001 ) .
.,width=302 ] we decided to use the h@xmath0 line for measuring radial velocities because : 1 .
the h@xmath0 region was observed on every night , providing a homogeneous dataset ; 2 .
this region is where the b9iii - type secondary ( orosz et al . 2001 ) and the a / f - type radial velocity standards have the most similar spectra , minimizing the effects of spectral mismatch on cross - correlation ( verschueren et al . 1999 ) ; 3 .
all the other stellar features are almost lost in continuum scatter .
we determined heliocentric radial velocities using 200 of the spectra , centred on the h@xmath0 line .
as the primary standard we chose hd 187691 , an f8v star , while @xmath12 vir ( f9v ) and hr 3383 ( a1v ) served as check stars for the absolute accuracy .
we preferred the f - type standard over hr 3383 because of its narrower h@xmath0 line , producing a slightly narrower cross - correlation peak .
a period search of the radial velocities resulted in @xmath18 d , which is in good agreement with the period determined by orosz et al .
( 2001 ) , and is practically identical to the ( more accurate ) photometric period .
for that reason , radial velocities were phased using the ephemeris in eq . 1 and are plotted in fig.[radvel ] .
to estimate the measurement error , we took the residual mean scatter of the sine - wave fit ( solid line in fig .
[ radvel ] ) and the scatter of the interstellar na i d velocities , which were also determined with cross - correlation .
both estimates gave consistent results ( @xmath19 km s@xmath2 ) and the symbol sizes in fig .
[ radvel ] reflect this uncertainty .
to check whether eq . 1 is consistent with the spectroscopic data , we also phased the data with the spectroscopic ephemeris obtained by orosz et al .
( 2001 ) : @xmath20 d , @xmath21 .
the resulting phase diagram was shifted by @xmath22 in respect to fig . 2 of orosz et al .
( 2001 ) , i.e. the minimum of the sine - wave occurred at @xmath23 .
we used this shift to correct the orosz et al .
period to @xmath24 d , indicating the good accuracy of the photometric period in eq . 1 . the best fit sine - wave in fig . [ radvel ] resulted in the following parameters : @xmath1 km s@xmath2 , @xmath25 km s@xmath2 .
the @xmath26 velocity amplitude is in excellent agreement with that of orosz et al .
( 2001 ) , who measured @xmath27 km s@xmath2 .
combined with the period , the resulting mass - function is @xmath28 m@xmath8 .
however , our @xmath29-velocity differs by over 30 km s@xmath2 ( the orosz et al .
value is @xmath30 km s@xmath2 ) , i.e. over 10@xmath31 . to determine the cause of the discrepancy , we first checked the consistency between different radial velocity standards .
vir , the other f - type standard , yielded essentially the same result , @xmath32 km s@xmath2 , while hr 3383 , the a1v standard with broader h@xmath0 line , gave @xmath33 km s@xmath2 ( the @xmath26 radial velocity amplitude was the same in all cases within the error bars ) .
note that the quoted errors came from the sine - wave fit , so that systematic errors are not included .
from these numbers , it is obvious , that _ ( i ) _ the large difference between the hd 187691 velocities and those of orosz et al .
( 2001 ) is real and _ ( ii ) _ despite the consistency of the two f - type standards , the hr 3383 velocities warn us that there might be larger systematic errors due to spectral mismatch than the raw data would suggest . in other words , it would be misleading to conclude from the two f - type standards that we have measured the @xmath29-velocity within @xmath34 km s@xmath2 . on the other hand ,
the broader h@xmath0 line of hr 3383 may enhance possible systematics in cross - correlation velocities ( for instance , due to uncertainties in continuum normalization ) , so that the apparent offset between the f - type and a - type standards ( @xmath167 km s@xmath2 ) is likely to be more affected by the a - type hr 3383 . considering these aspects
, we think the simple mean of the three values and its standard deviation give a more realistic @xmath29-velocity and error estimate than , for instance , the average of the two f - type velocities .
we adopted therefore @xmath3 km s@xmath2 as our systemic velocity of v4641 sgr .
we then noticed that in sect . 2 of orosz et al .
( 2001 ) its authors used a bright night sky emission line to match the wavelength scales of spectra taken in different observatories : _ `` the bright night sky emission line at @xmath355578 was used to make small adjustments to the wavelength of the flwo and vlt spectra .
the shifts required to align this feature to a common wavelength ( 5578.0 ) were generally less than @xmath3515 km s@xmath2 , although the vlt spectra from the end of the night of june 7 required shifts on the order of 80 km s@xmath2 . ''
_ we checked this line and found that it belongs to the neutral oxygen , for which the nist atomic spectra database lists 5577.339 as the laboratory wavelength . on the other hand ,
iraf s internal telluric line list ( linelists / skylines.dat ) contains two telluric features at 5577.3894 and 5577.5948 .
these three wavelengths are @xmath36 km s@xmath2 towards blue from the value used by orosz et al .
( 2001 ) , which suggests that the orosz et al .
data are offset by about 30 km s@xmath2 . if we adopt the o i line from the nist database , the orosz et al .
data should be corrected by 35.5 km s@xmath2 , yielding a corrected @xmath29-velocity of @xmath37 km s@xmath2 , which is in perfect agreement with ours .
this is very reassuring , as the @xmath38 km s@xmath2 shift in the centre - of - mass velocity over @xmath39 years could otherwise be explained only by a very massive invisible third body .
this not only seems very unlikely , given the @xmath40 m@xmath8 total mass of the binary ( orosz et al .
2001 ) , but also lacks support from other observations . finally , we present the trailed quiescence spectra of v4641 sgr in fig . [
trailed ] , in which we emphasize the weaker lines of the secondary star and the lack of spectral changes other than varying doppler - shifts .
besides h@xmath0 , five further lines ( of o , mg and si ) clearly showed the orbital motion .
other lines remained constant ( except for a @xmath1630 km s@xmath2 scatter in the telluric lines due to the heliocentric corrections calculated with the _ rvcor _ and _ dopcor _ tasks ) .
the stability of these spectra indicated a minimum level of mass - transfer activity during our observations .
in black - hole x - ray binaries , emission lines are generally from an accretion disk .
they have double - peak structures , typically with an equivalent width of 10 - 20 in h@xmath0 ( see a recent discussion in wu et al .
additionally , their line profiles often have weak asymmetry due to orbital motion of a bright spot ( e.g. casares et al . 1995 ) or extended structures around the accretion disk ( soria et al . 2000 ) .
the emission lines of v4641 sgr are completely different from these tyical results .
firstly , the lines are highly asymmetric , a feature which is difficult to attribute to the weak contribution of a bright spot .
secondly , the equivalent width of the h@xmath0 is quite large , 130170 .
the emitting area of lines should therefore have a volume at least five times that of the entire disk which argues against scenarios with local asymmetry in a disk .
furthermore , both fe and he lines show p cygni profiles , which are not expected from an accretion disk .
consequently , instead of emission from a disk , the evidence points to a violent mass ejection and the presence of an outflow , which began optically thick in the paschen lines and became optically thin within five hours .
this suggests that the ejection was very limited in time and our observations covered the initial phase of shell expansion .
although it has a very different character , we noticed a remarkable spectral similarity with the x - ray transient ci cam ( rxte j0421 + 560 ) .
robinson et al . ( 2002 ) presented h@xmath0 and paschen line profiles of the 1998 march / april outburst of ci cam , which had the same broad blue extensions like v4641 sgr in outburst .
hynes et al .
( 2002 ) , whilst discussing their optical spectra which showed similar line profiles to ours , proposed a two - component outflow for ci cam , with a fast , hot polar wind and a dense , cooler equatorial outflow with much lower velocity . however , the temperature of the secondary star in v4641 sgr is too low to generate a strong stellar wind , so that the outflow is presumably from a disk , in other words , a disk wind .
regarding this point , we found another noteworthy similarity , this time between v4641 sgr and the famous galactic jet source , ss 433 .
spectra taken from it show an h@xmath0 emission line with both `` moving '' components , which are interpreted as optical jets , and a `` stationary '' component ( for examples of the latter , see murdin et al .
( 1980 ) , vermeulen et al . (
1993 ) and gies et al .
( 2002 ) ) .
the profiles of this stationary line of ss 433 depend on its orbital phase .
consequently , it is very interesting that the emission line of v4641 sgr , which was taken near an orbital phase of 0 , is quite similar to the stationary component of ss 433 when its orbital phase = 0 .
given that the stationary component in ss 433 is interpreted as originating from a disk wind , this provides further evidence for the same in v4641 sgr .
as both systems have a high inclination angle , the presence of a secondary star probably affects line profiles since it would block a part of the wind at an orbital phase = 0 .
this may cause a weaker blue component in h@xmath0 and he lines of v4641 sgr .
inspired by the similarity with ss 433 , we unsuccessfully tried to find possible `` moving emission lines '' for a sign of optical jets around the h@xmath0 line of v4641 sgr .
a possible feature can be seen at about 6360 , but this is more likely to be due to atmospheric distortion of the mg / si ii emissions .
however , our spectral coverage was quite limited and we can not firmly exclude the existence of `` moving emission lines '' outside the 60477005 region . from a theoretical point of view
, disk winds have been studied for cataclysmic variables , especially novalike systems .
for example , proga ( 2003b ) calculated resonance - line profiles predicted by radiation - driven disk wind models for novalike systems .
some of his theoretical emission line profiles are similar to lines in v4641 sgr and ss 433 . however , despite the recent progress in the field ( e.g. feldmeier & shlosman 1999 , feldmeier et al .
1999 , proga 2003a ) , the complexity of the problem prevents simple modeling of the observed line profiles .
an important point is that this line - driven wind theory needs a strong uv radiation to accelerate the wind .
the strong uv radiation can be expected from the surface of white dwarfs in cataclysmic variables . on the other hand ,
the uv radiation from black hole x - ray binaries is expected to be weaker , especially in a low / hard state of x - ray binaries .
this is because a hot thermal accretion disk is truncated by an inner `` radiation inefficient accretion flow '' ( riaf ) .
the most fundamental difference between v4641 sgr and ss 433 is their x - ray luminosities .
the x - ray luminosity of ss 433 reaches super - eddington luminosity for a 10 m@xmath8 black hole ( @xmath1610@xmath41 ergs s@xmath2 ) . according to theories of super - eddington accretion , the accreting gas forms an optically thick and geometrically thin disk ( so called `` slim disk '' , abramowicz et al .
. this slim disk can generate strong uv radiation , which is favourable for the line - driven wind model .
but , the x - ray luminosity of v4641 sgr is about @xmath1610@xmath42 erg s@xmath2 during minor outbursts .
this value indicates a low / hard state , in which we can not expect strong uv radiation . in summary ,
the driving mechanism of the disk wind in v4641 sgr remains an open issue .
it is possible that the driving mechanism is not the uv radiation , but a magnetic field .
another possibility is that the accretion rate actually reached the super - eddington level , but was undetected perhaps because most of the x - ray radiation was absorbed by a thick torus around the black hole and reemitted in optical .
currently though , both scenarios are merely speculations .
another important question is : to what extent was the 1999 `` super - outburst '' different from the smaller eruptions in 2002 , 2003 and 2004 ?
based on the orders - of - magnitude lower x - ray emission in the later outbursts , uemura et al .
( 2004a , 2005 ) concluded that all minor outbursts had the same nature , which was different from that of the 1999 september event .
uemura et al .
( 2004a ) also proposed a scenario for the time evolution of the rapid brightness fluctuations : generation of a hot region in an accretion disk , which propagates towards the inner portion of the disk , triggering short - term fluctuations .
when the active and bright region finally disappears , which is observed as a dip , the whole cycle repeats with the replenishment of gas from the outer region .
our optical and near - infrared spectra suggest that despite the much weaker x - ray flux , mass ejection had very similar kinematic properties in 1999 september and 2004 july . most noticeably , the description of the h@xmath0 line in the 1999 outburst by orosz et al .
( 2001 ) and chaty et al .
( 2003 ) fits our observations in 2004 perfectly .
the broad , strong emission lines and their variability ( e.g. the he i @xmath155015 and he i @xmath155049 lines ) suggest the presence of a high - velocity outflow component blown off from the accretion of matter on to the compact object .
( schulz & brandt 2002 , chaty et al . 2003 ) .
the paschen lines revealed the surprisingly short time - scale of this blow - off .
the first spectra contained strong emission in the upper lines , which is indicative of high density / optically thick lines ( e.g. lynch et al .
2000 ) . within five hours
, upper paschen lines strongly decreased , implying switching to optically thin emission , so that the blown - off matter did not have further supplies from the accretion disk .
the variations of the h@xmath0 equivalent width during a 10-min flare - up revealed that rapid brightness fluctuations are caused by continuum variations , therefore our observations are consistent with the uemura et al .
( 2004a ) scenario in a sense that the source of these continuum fluctuations must be in the inner part of the accretion disk .
we also found subtle changes in the h@xmath0 line profile , which suggest the presence of additional phenomena during the rapid fluctuations : whereas the redshifted emission component did not show any correlation with the flare - up , the blue and the broad components exhibited significant changes in the velocity range of the emitting gas .
moreover , the gradual shifts with opposite signs in the line centre of the blue and red components support the contention that the three components ( blue , red and broad ) formed in well - separated parts of the system . the source of the red component must have been located far out on the accretion disk , as it was left unchanged by the flare . in contrast , the blue and broad components traced the continuum changes very well , suggesting they must have originated much closer to the disk , with the broad component being the closest .
the outburst h@xmath0 spectra allowed the detection of the 6613 diffuse interstellar band ( dib ) on the red wing of the h@xmath0 profile .
the equivalent width of this dib is known to correlate with the interstellar reddening ; adopting @xmath43 ( jenniskens & dsert 1994 ) , the observed @xmath44 yields @xmath45 mag .
the sodium d doublet in quiescence is significantly stronger than reported for the 1999 outburst by chaty et al .
( 2003 ) : we measured the equivalent width of the sodium d@xmath46 line between 0.8 0.9 in the best quiescence spectra , which is about twice as much as the 0.45 observed by chaty et al .
. however , our spectra have reasonably better resolving power ( @xmath47 vs. @xmath48 ) and the given @xmath49 suggests slightly higher reddening , between 0.40.6 mag ( see fig . 4 in munari & zwitter 1997 ) .
considering the large uncertainties , these reddenings are consistent with the previous determinations : @xmath50 mag ( orosz et al .
2001 ) and @xmath51 mag ( chaty et al .
2003 ) .
our radial velocity curve has the smallest uncertainties in the literature so far ; the @xmath26 velocity amplitude ( @xmath52 km s@xmath2 ) and the corresponding mass - function ( @xmath53 m@xmath8 ) are the same as those of by orosz et al .
the most significant improvement is found for the centre - of - mass velocity of the system , for which we adopt the mean value of the three independent measurements : @xmath3 km s@xmath2 .
assuming that this systemic velocity is due entirely to differential galactic rotation , we slightly revise the kinematical distance limit to @xmath54 kpc , using the rotation curve given in fich et al .
( 1989 ) and the standard iau rotation constants of @xmath55 kpc and @xmath56 km s@xmath2 ( from @xmath57 kpc , orosz et al .
when combined with accurate distance and proper motion measurements , the @xmath29-velocity will be important in finding the galactocentric orbit of v4641 sgr .
knowing the galactic motion of an x - ray binary can help understand the possible origin of the system , which in turn can reveal intriguing details of star formation in the early stages of evolution of the galaxy ( mirabel et al .
2001 , 2002 , mirabel & rodrigues 2003 ) .
we have presented optical and near - infrared spectra of the galactic microquasar v4641 sgr in a short outburt in 2004 july and quiescence between 2004 september2005 march .
the conclusions can be summarized as follows : 1 .
the photometric behaviour of the 2004 july outburst was similar to those in 2002 and 2003 , i.e. the star showed rapid brightness fluctuations , due to changes in the continuum radiation .
however , the complex multi - component emission line profiles of hydrogen , helium and iron were similar to those observed in the big outburst of 1999 september , with evidence of a similar high - velocity outflow and optically thick and then thin emission .
this suggests that kinematic properties of the mass - ejection processes were similar in 1999 and 2004 despite the huge difference in peak luminosities of the eruptions . 2 .
we have determined the most accurate radial velocity curve of the system so far .
we confirm the optical mass function of the system by orosz et al .
( 2001 ) reducing its uncertainty by a factor of three to @xmath28 m@xmath8 .
we correct the systemic velocity by more than 30 km s@xmath2 to @xmath3 km s@xmath2 .
we do not find spectral changes other than the doppler - shifts due to the orbital motion , so any accretion activity must have been at low levels during our quiescence observations .
3 . in the last couple of years
it has been clearly established that v4641 sgr has a recurrent activity with a time - scale of 12 years .
looking back , the shorter value seems to be more appropriate , which means every observing season should be monitored for further eruptions .
our findings demonstrate the importance of taking rapid spectroscopic snapshots right after the discovery of a new outburst .
we suggest to keep the object on lists of possible target - of - opportunity observations , because collecting more information on spectral changes in early phases is expected to yield better understanding of interactions around the black holes in systems like v4641 sgr , especially the nature of mass ejection and its connection to the accretion disk .
this work has been supported by the otka grant # t042509 and the australian research council . c. lindstrm received a vacation scholarship for this project from the school of physics , university of sydney .
kiss is supported by a university of sydney postdoctoral research fellowship .
a. derekas joined the project within the international postgraduate research scholarship scheme of the australian department of education , science and training .
the authors thank prof .
hunstead for a careful reading of the manuscript .
the nasa ads abstract service was used to access data and references .
wu k. , soria r. , johnston h. , hunstead r. , 2002 , in : proc .
of the ninth marcel grossmann meeting , eds .
: vahe g. gurzadyan , robert t. jantzen , remo ruffini , world scientific publishing , p. 2274
( full text as astro - ph/0101390 ) | we present spectroscopic observations of the black - hole binary v4641 sagittarii , obtained between 4th july 2004 and 28th march 2005 , which cover the minor outburst of the star in early july 2004 and quiescence variations on 19 nights scattered over six months . during the outburst ,
the star peaked approximately 3 magnitudes brighter than usual , and our spectra were dominated by broad hydrogen , helium and iron emission lines .
the very first spectra showed p cygni profiles , which disappeared within a few hours , indicating rapid changes in matter ejection .
the h@xmath0 line had multiple components , one being a broad blue - shifted wing exceeding 5000 km / s . during a simultaneously observed 10-min photometric flare - up
, the equivalent width of the h@xmath0 line temporarily decreased , implying that it was a flare of the continuum .
the overall spectral appearance was similar to that observed in the 1999 september active phase , which suggests that similar mass - ejection processes were associated with both eruptions . in quiescence ,
the spectra were those of the early - type secondary star showing its orbital motion around the primary . by measuring cross - correlation radial velocities ,
we give an improved set of spectroscopic elements . whereas we measure the same velocity amplitude ( @xmath1 km s@xmath2 ) , within errors , as orosz et al .
( 2001 ) , our centre - of - mass velocity ( @xmath3 km s@xmath2 ) differs significantly from the previously published value ( @xmath4 km s@xmath2 ) .
however , we find evidence that the difference is caused by a systematic error in data reduction in the previous study , rather than by gravitational effects of an invisible third component .
x - rays : binaries black hole physics stars :
individual : v4641 sgr |
a well known characteristic of accreting stellar mass black holes is the presence of two drastically different spectral states , first discovered in cygnus x-1 by tananbaum et al .
1972 ( see tanaka and shibazaki 1996 for review ) . in
the _ soft _ spectral state luminosity peaks at around 1 kev , while in the _ hard _ state luminosity is dominated by photons with the energy of the order of 100 kev .
this bimodality is believed to be related to very different regimes of the accretion flow in the vicinity of the black hole .
soft radiation is interpreted as a multicolor black body emission originating from the optically thick ( geometrically thin ) disk ( shakura , sunyaev 1973 ) , while hard emission , having a nearly power law shape at low energies and a cutoff above @xmath0 kev , should come from an optically thin and hot medium where comptonization of soft seed photons by the hot electrons plays an important role ( e.g. sunyaev and trmper 1979 ) .
one of the popular models assumes that an optically thick disk is truncated at some distance from the black hole and followed by an optically thin and hot flow ( see e.g. thorne and price 1975 , liang and price 1977 , esin , mcclintock and narayan 1997 , meyer , liu and meyer - hofmeister 2000 ) .
the hard and soft spectral states of the source may then correspond to the situation when an optically thick disk is truncated far from or close to the black hole respectively .
another important property of the x ray emission from black hole candidates is a strong aperiodic variability on time scales longer than 10 ms ( see e.g. van der klis 1994 ) .
the power density spectra are very different during different spectral states suggesting that the variability and spectral properties are closely linked .
cygnus x1 is the best studied accreting black hole in our galaxy .
due to its brightness and persistent nature it was observed virtually by every x ray observatory flown to date .
the source was observed in different spectral states , also having very different properties of the short time scale variability .
we discuss below qualitative model aimed to explain the changes in short time scale variability correlated with changes in the spectral shape .
the state of the black hole candidates , when the spectrum contains a strong soft component , are usually called `` high / soft '' or `` very high '' states ( e.g. tanaka and shibazaki 1996 ) .
this soft spectral component has a shape resembling multicolor black body emission and is believed to be produced by the standard optically thick ( geometrically thin ) disk of shakura and sunyaev ( 1973 ) type .
along with this `` black body '' emission a harder component is often present in the spectrum , with an approximately power law spectrum .
study of the variability properties of these two components led to the conclusion that most of the variability is associated with the power law component and not with the `` black body '' emission ( see e.g. miyamoto et al .
1994 for the ginga observations of nova musca 1991 ) .
the same behavior have been observed for various sources manifested by the increase of the fractional rms with energy . for the rxte @xcite observations of cygnus x1 in the soft state in june 1996 gilfanov , churazov and revnivtsev ( 2000 ) found that on the time scales shorter than 100 seconds the amplitude of the soft component variations is at least an order of magnitude lower than that of the harder component . for illustration
we plot in fig.[cor]a the pca / rxte count rate in the soft band @xmath1 as a function of the count rate in the harder band @xmath2 .
here @xmath3 is the observed ( background subtracted ) light curve in a given energy band .
the data points are the 16 s averaged values of the count rate from the `` standard mode 2 '' format of pca .
the observations were performed in 1996 on june 4 and 16 . from fig.[cor]a
one can see that ( * i * ) relative amplitude of variations in the soft band is a factor of @xmath4 smaller than in the hard band and ( * ii * ) the correlation can be reasonably well approximated by a linear relation @xmath5 , where @xmath6 .
this relation is shown in fig.[cor]a with a straight line .
similarly good linear relation exists between the count rates in any pair of the pca / rxte channels . should the observed spectrum vary in intensity only and not in shape
, then one would expect linear relation and @xmath7 .
the fact that the relation between the count rates in two energy bands is close to linear , but @xmath8 is not zero , means that additional stable component contributes to one or both energy bands .
for any two energy bands above @xmath9 kev the count rate in one band is approximately proportional to the count rate in another band , i.e. @xmath8 is close to zero .
very significant deviations of @xmath8 from zero level appear if one of the selected energy bands is below 5 kev .
this suggests that stable component is present mostly at low energies .
we then repeat the same procedure of fitting the linear relation between @xmath10 and @xmath11 for every energy channel setting @xmath10 to the count rate in this channel and always choosing @xmath11 as a count rate above 9 kev .
thus @xmath12 , where @xmath13 and @xmath14 are the observed light curves .
the @xmath11 therefore serves as a `` reference '' light curve while @xmath15 and @xmath16 characterize respectively the contribution of the stable component to a given channel and the coefficient of proportionality between the variations of the count rate in this channel and the variations of the count rate above 9 kev . for the energy channels above 9 kev
there should be additional intrinsic correlation between @xmath17 and @xmath11 because @xmath11 contains contribution from @xmath17 , but experiments with various choices of the reference light curve @xmath11 ( e.g. using only narrow energy range instead of all counts above 9 kev ) proved that resulting @xmath15 and @xmath16 vectors are relatively insensitive to the choice of @xmath11 as long as @xmath11 contains only contributions from energy channels above @xmath18 5 kev . the resulting vectors @xmath15 and @xmath16 may be interpreted as the spectra ( in units of @xmath19 ) of stable and variable components . if the stable component is indeed present only at low energies then with our choice of the @xmath14 the @xmath15 will reproduce both the shape and the normalization of the stable component , while @xmath16 will recover only the shape of the variable component .
the normalization of @xmath16 depends on the particular choice of the `` reference '' light curve @xmath11 .
these @xmath15 and @xmath16 spectra have been approximately unfolded using the xspec v10 @xcite command * eeufspec * and assuming a power law model with a photon index of @xmath182.5 .
this procedure simply divides the observed count rate in each channel by the effective area in this channel , calculated for the assumed spectral model .
the photon index of 2.5 approximately corresponds to the shape of the power law tail in the source spectrum during june 1996 observations ( gierliski et al . 1999 ,
gilfanov et al .
the spectra of constant and variable components unfolded that way are shown in fig.[spec]b .
the stable component @xmath15 has a very soft spectrum ( open circles in fig.[spec]b ) . for comparison
the light grey line shows the spectrum of the multicolor black body disk emission with the characteristic temperature @xmath20 kev .
the `` variable '' component @xmath16 has a much harder spectrum ( filled circles in fig.[spec]b ) and does not contain a strong soft component ( see gilfanov et al .
2000 for details ) .
note that although a good linear relation between count rates in any pair of energy channels means that deconvolution in two components is accurate enough , these two spectral components may not necessarily have direct physical meaning .
however the fact that stable component closely reproduces the multicolor black body disk emission both in terms of shape and normalization ( gilfanov et al .
2000 ) strongly suggests that stable component obtained from the above analysis simply coincides with the multicolor black body disk emission .
finally the two upper spectra shown in fig.[spec]b with solid squares were obtained averaging the observed spectra over the periods of time when the count rate above 9 kev was high and low respectively .
these spectra contain data points scattered over a period of time of more than 10 days when intensity of the source in the hard band vary by at least a factor of @xmath184 ( see fig .
the thick grey lines show that these spectra can be reasonably well ( within 1015% ) approximated by a model @xmath21 consisting from the above obtained stable and variable spectral components where @xmath22 ( the normalization of the variable component ) is the only free parameter .
thus from this analysis one can conclude that all spectra observed by rxte on 1996 , june 4,16 during individual 16 seconds exposures can be reasonably well approximated by a combination of a constant soft component and a harder component which vary strongly in amplitude , but not much in shape . the typical power density spectrum ( pds ) of the x ray flux from cygnus x-1 ( in the 613 kev energy range ) is shown in fig.[pdssoft ] .
the pds is calculated in units of squared fractional rms per hz , corrected for the poissonian noise contribution . in softer energy bands the pds has similar shape , but lower normalization , because of the larger contribution of the stable soft component to the source flux . the energy spectrum of cygnus x-1 in the soft state is discussed in details in gierliski et al .
1999 ( see also dotani et al . , 1997 and zhang et al . , 1997 ) .
the strong soft component was present in the cygnus x-1 spectrum during these observations ( even in the rxte band above @xmath183 kev ) .
the characteristic temperature of the black body component is @xmath180.5 kev
( gierliski et al . 1999 ) and this component provides more luminosity than the power law tail .
assuming that the black body component is due to an optically thick disk and that no strong advection is present in the optically thin flow one can conclude that the inner radius of the disk is rather close to the black hole .
accurate determination of the disk inner radius from the 1996 june 4,16 rxte spectral data is difficult because of the poor rxte spectral response below 3 kev , but the analysis of simultaneous rxte and asca 1996 may 30 data ( dotani et al .
1997 , gierliski et al . 1999 ) indeed indicates that the inner radius of the disk is close to the black hole .
we make a conservative assumption that the inner edge of the accretion disk @xmath23 is within @xmath24 , where most of the gravitational energy of the accreting matter is released .
the pds of x ray flux variations in this state ( fig.[pdssoft ] ) holds the same shape ( @xmath25 flicker noise ) from few @xmath26 hz up to 10 hz , suggesting that the same physical mechanism is responsible for the flux variations at all frequencies in this range .
since these variations are associated with the power law spectral component they presumably originate from an optically thin region . taking into account that @xmath25 slope of the pds holds down to the frequencies of few @xmath26 hz it is difficult to imaging that observed flux variations over this extremely broad dynamic range of time scales ( at least 45 orders of magnitude )
can be provided by instabilities developing in the innermost region within @xmath23 .
more promising seems the assumption that the observed variability is due to instabilities occurring at different distances ( much larger than @xmath23 ) from the black hole and then propagating into the innermost region , where the energy is released and x ray photons are produced . in particular lyubarskii ( 1997 ) considered the fluctuations of mass accretion rate in the innermost region associated with the fluctuations of the viscosity parameter @xmath27 in the accretion flow at much larger radii .
if the amplitude of @xmath27 fluctuations at any radius is the same then variations of the mass accretion rate through the boundary , placed at much smaller radius , will have an @xmath25 power density spectrum .
thus the mass accretion rate is modulated at large distances from the black hole , but the observed ( modulated ) x ray flux is coming from the innermost region .
the broad dynamic range of the variability time scales in the model of lyubarskii ( 1997 ) can be provided by the broad range of radii at which the viscosity is fluctuating .
thus on one hand a stable optically thick disk extends down to a very small radii @xmath23 ( as indicated by the strong and stable soft component ) and on the other hand prominent variations of the harder component are present in a broad range of time scales ( up to at least @xmath28@xmath29 s see fig.[pdssoft ] ) , which are likely to be associated with instabilities occurring at much larger radii than @xmath23 .
the simplest explanation required to combine these two facts together is the assumption that along with a stable optically thick accretion disk a variable ( optically thin ) corona is present , which extends in radial direction up to large distances from the black hole .
the large extent of the corona in radial direction is required in order to provide broad dynamic range of the variability time scales .
various models involving optically thin corona have been discussed in the literature ( e.g. bisnovatyi kogan and blinnikov 1976 , liang and price 1977 , galeev , rosner and vaiana 1979 , haardt and maraschi 1991 , esin , mcclintock and narayan 1997 , esin et al .
1998 , gierliski et al . 1999 , see poutanen 1998 for recent review ) .
the configuration of the accretion flow we adopted is schematically shown in fig.[sk1]a . here
the solid black slab shows the optically thick accretion disk , which is assumed to be stable , sandwiched by an optically thin corona which is shown as a grey shaded region .
it is further assumed that instabilities are operating in the corona which modulates the mass accretion rate flowing through given radius .
the `` sine wave '' with varying period schematically shows that the time scales of modulations in the coronal flow increase with the radius .
these modulations of the mass accretion rate are propagated down to the inner region of the main energy release ( shown by the thin box near the black hole ) where the observed x ray emission is produced .
thus the variations of the observed x ray flux coming from the inner region of the main energy release reflect the modulations of the mass accretion rate at much larger radii . in this picture
one would then expect to observe the spectrum consisting of two components ( soft stable component due to the disk emission and harder variable component due to comptonization in the corona ) .
the relative contribution of these two components to the luminosity of an averaged spectrum would then reflect the ratio of the energy releases in the disk and corona ( or mass accretion rates if no strong advection takes place in the corona ) .
the pds of the harder component may then have the same shape over broad range of frequencies as indeed observed ( fig.[pdssoft ] ) .
the relative amplitude of the x ray flux variations should be lower at low energies where stable multicolor black body component provides dominant contribution the observed flux .
let us now consider what kind of pds we can expect in the hard state . in the hard state the soft ( black body )
component is weak or absent .
we assume below that that the optically thick disk is truncated at a larger distance from the black hole ( so that emission from the disk falls below x ray regime ) .
the adopted configuration of the accretion flow is schematically shown in fig.[sk1]a . here
the optically thick ( stable ) disk ends at some radius and it is followed by an optically thin flow which joins the coronal flow , which is present above the disk .
we assume that properties of this inner optically thin flow in terms of amplitude and characteristic time scales of modulations of the mass accretion rate are similar to those of the corona .
we further assume that above the disk truncation radius larger fraction of the mass accretion rate @xmath30 is flowing through the thin disk rather then through the corona @xmath31 .
i.e. @xmath32 .
the total mass accretion rate in the optically thin flow below the disk truncation radius is @xmath33 .
the pds ( the black solid line in fig.[sk2]b ) is then expected to have several distinct regions over frequency .
these regions are schematically shown by the vertical lines in the fig.[sk2]b labeled as @xmath34 , @xmath35 , @xmath36 .
the corresponding geometrical regions in the accretion flow are marked with the arrows in fig.[sk1]a . at high frequencies ( @xmath37 in fig.[sk2]b )
there is a turnover of the pds which may be due to the same reason as the turnover in the soft state pds ( fig.[pdssoft ] ) .
the discussion of this turnover is beyond the scope of this paper and we only briefly speculate on it in section 3 .
detailed analysis of the high frequency part of the cygnus x1 pds is given in revnivtsev et al .
, 2000 .
the range of frequencies from @xmath35 through @xmath36 is associated with the variability scales corresponding to the optically thin region of the flow below the disk truncation radius ( fig.[sk1]a ) . in this region we might expect the flow to be similar to the corona flow in the soft state and as a result the pds should roughly reproduce the pds in the soft state ( i.e. @xmath38 for @xmath39 ) . the only difference is that in the soft state additional soft component is contribution to the observed flux .
this soft component does not contribute much at the energies higher than 5 kev and we therefore can directly compare the shape and the normalization of the pds in the hard and soft state for the higher energy bands ( fig.[pds]c ) . here
@xmath35 is a characteristic frequency of the modulations introduced by instabilities operating in the optically thin flow at the truncation radius of the disk . below this frequency
only a small fraction of accreting matter contributes to variability the fraction of matter which goes through the corona . as a result normalization of
the power density spectrum , associated with the viscosity fluctuations in this region should drop by a factor of @xmath40 .
this component should appear in the power density spectra as a @xmath25 component at very low frequencies as shown in fig.[sk2]b .
the frequency below which this component is dominating the pds is marked as @xmath34 .
there is also a transition region between frequencies @xmath34 and @xmath35 .
one can expect the pds to be flat in this frequency range .
indeed , the variability of the x ray flux in this range of time scales is not associated with the `` locally '' induced fluctuations of the accretion flow , which are suppressed by the factor @xmath41 , but rather with the stochastic superposition of the modulations of the accretion rate occurring at smaller radii .
the frequency @xmath34 , below which the pds switches from flat to @xmath25 dependence can then be expresses as @xmath42 .
thus the assumption of the variable corona sandwiching the stable disk ( as inspired by the soft state data fig.[pdssoft ] ) leads to a prediction of a specific shape of the pds in the hard state .
for comparison observed pds in the hard state ( for two rxte observation in march 1996 ) are shown in fig.[pds]c . using the above arguments on the relative normalization of the two @xmath25 regions of the pds
one can estimate that in cygnus x-1 the fraction of mass accretion rate going through the corona above the disk is @xmath18 20% .
variations in the pds shape ( in particular the shift of the break frequency ) is then interpreted as the change of the disk truncation radius and the related change of the characteristic frequency .
the expected change of the pds caused by inward shift of the disk radius is shown by the dashed line in fig.[sk2]b .
similar behavior of the pds ( correlated change of the break frequency and normalization of the band limited noise ) was first reported for cyg x-1 by belloni & hasinger ( 1990 ) .
when the disk truncation radius extends well down to the innermost region the pds switches from the `` 3-breaks '' shape to the `` 1-break '' shape as observed in the soft state .
described above is a qualitative picture inspired by the variability of the source during soft state . the suggested model is a phenomenological one and we speculate below on the possible underlying physics .
the first important question is why the mass accretion rate in the thin disk would be stable , while in the corona it is variable .
note that we do not address here the question of the accretion flow stability ( e.g. against thermal and viscous instabilities in the radiation - pressure - dominated part of the disc ) .
instead one can think of the effect on the mass accretion rate of the magneto - hydrodynamic turbulence , which presumably serves as a source of the viscosity in the accretion flow through the fluctuating magnetic stresses ( e.g. hawley , gammie and balbus 1995 , brandenburg , nordlund , stein and torkelsson 1995 ) .
lyubarskii ( 1997 ) considered the power density spectrum arising from fluctuations of viscosity at different radii which causes fluctuations of the mass accretion rate .
fluctuations at one radius are related to the fluctuations at smaller radii through the green function of the diffusion equation ( lynden bell & pringle 1974 , lyubarskii 1997 ) . in his picture fluctuations of viscosity at a given radius on the viscous time scales @xmath43 causes fluctuations of the mass accretion rate at all smaller radii . here
@xmath44 is a keplerian angular frequency at a given radius @xmath45 , @xmath27 is viscosity parameter of shakura & sunyaev 1973 , @xmath46 is the half thickness of the disk .
however because of the diffusive nature of the disk accretion any fluctuations of the mass accretion rate on time scales much shorter than the diffusion time scale will not propagate towards much lower radii ( see appendix ) , but instead will vanish in amplitude very close to the radius at which they originated .
e.g. if we assume that actual fluctuations at a given radius are occurring at a time scales @xmath47 , comparable with the dynamical time scales @xmath48 or thermal time scales @xmath49 , then in the standard thin disk , where @xmath50 , we will always have @xmath51 .
therefore such fluctuations will never propagate down to much smaller radii . even if we consider the fluctuations occurring in the inner zone of the geometrically thin disk , which is emitting in x
rays ( i.e. multicolor black body component directly observed by x ray telescopes ) we can easily show that the amplitude of fluctuations on a time scale @xmath52 will be significantly suppressed after propagating a distance @xmath53 in a radial direction such as : @xmath54 . for
larger @xmath55 the amplitude of the mass accretion rate fluctuations will vanish .
e.g. for fluctuations on the thermal scales @xmath56 ( @xmath57 is a typical value for the standard geometrically thin disk , dominated by a gas pressure ) .
therefore @xmath58 different ( `` incoherent '' ) region over radius will contribute to the observed flux , effectively suppressing the fluctuations by a factor of @xmath59 .
thus for the geometrically thin disk fluctuations of @xmath27 ( or mass accretion rate ) on the dynamical of thermal time scales will not cause very prominent variations in the observed flux . on the other hand
geometrically thick disks ( e.g. adaf flows ) are much more transparent for the high frequency oscillations .
e.g. the fluctuation of the viscosity ( @xmath27 parameter ) at some radius ( at the dynamical or thermal time scales ) will affect the mass accretion rate at all smaller radii and thus may provide fluctuations of observed flux coming from the innermost region of the accretion disk . thus qualitatively the apparent stability of the disk compared to the corona could be understood as the result of a much longer diffusion time in the disk , which suppresses propagation of fluctuations . in the above model
the characteristic frequencies are related to the position of the disk truncation radius : the smaller the truncation radius the higher the characteristic frequencies ( in particular the break frequency ) .
at least qualitatively this trend is consistent with correlation of the spectral and timing parameters observed in the black hole candidates .
e.g. for cygnus x1 and gx 3394 the increase of the characteristic frequencies correlates with the steepening of the spectra and the increase of the reflected component ( gilfanov et al . , 1999 , revnivtsev et al . , 2000 ) , which may be related to the increase of the cooling of the comptonization region by the soft photons from the disk and the increase of the fraction of the hard flux intercepted by the thin disk as it approaches black hole . in grs1915 + 105 and xte
j1550564 the qpo frequency , which in turn correlates with the break frequency , is well correlated with the soft component flux ( e.g. trudolyubov et al . 1999 , muno et al . 1999 ) . of course the above representation of the pds as
a 3-break function is a gross oversimplification . in reality
the shape of the pds will be much more complex .
e.g. broad humps may appear near the break frequency because the geometrically thin disk at the truncation radius supplies mass at a steady rate to the `` unstable '' inner geometrically thick region . in such conditions any increase of the accretion rate in the optically thin region must be followed by the decrease of the accretion rate at later moments of times ( since total mass supply rate by the thin disk is constant ) . as a result a broad qpo
like hump may appear in the power density spectrum ( vikhlinin , churazov , gilfanov , 1994 ) .
the truncation radius of the disk may fluctuate with time and affect the observed x ray flux .
therefore even if the thin disk does not propagate the fluctuations of the mass accretion rate ( at the frequencies higher than diffusion time ) we may see fluctuations of the soft flux due to fluctuations of the disk truncation radius .
the only case when we should expect the soft component to be very stable is when the disk extends all the way down to the marginally stable orbit at @xmath60 ( i.e. in the genuine soft state ) . in the `` transition '' state
( i.e. when the disk is close enough to the black hole ) fluctuations of the soft component due to disk truncation radius can be observed directly . in the hard state (
when the disk is presumably truncated at a large distance from the compact source ) emission of the thin disk is almost outside the x ray band and variations of the soft disk flux can be observed indirectly through the influence of the soft flux on the comptonization region .
e.g. variations of the soft flux entering the comptonization region with a given optical depth and energy release to the electrons will result in variations of both the flux and slope of the comptonized spectrum .
the high frequency turnover of the pds ( at about 10 hz ) may be related to the instabilities operating in the region of the main energy release .
part of the gravitational energy has already been released ( and emitted ) at larger radii .
therefore the amplitude of x ray flux variations associated with the mass accretion rate fluctuations added to the flow in the inner region may be suppressed .
note that in the case of accretion onto a neutron star a large fraction of energy is released at the neutron star surface .
therefore this turnover may be absent in the power density spectra of the accreting neutron stars .
in the soft spectral state of cygnus x-1 , observed by rxte in june 1996 , the black body component was remarkably stable while the harder ( power law like ) component varied strongly on the frequency scales from 10 hz down to few @xmath26 hz .
we suggest that such behavior is due to presence of an optically thin corona above the optically thick disk , which extends up to a large distance from the black hole .
the variations of the mass accretion rate ( or viscosity ) occurring at a large distance from the compact object are propagated down to the region of the main energy release .
the reason for different variability properties of the disk and corona ( namely stable disk and unstable corona ) may be due to the fact that in the optically thick ( geometrically thin ) disk any fluctuations at time scales shorter than the diffusion time scales @xmath43 are effectively dampened and are not propagated down to small radii .
the assumption that in the hard state the disk is truncated at some distance from the black hole ( larger than the last marginally stable orbit ) then naturally lead to an explanation of the overall shape of the power density spectra of black hole candidates in this spectral state .
we are grateful to philip armitage , yura lyubaskii , friedrich meyer , henk spruit and rashid sunyaev for useful discussions , andrzej zdziarski and anonymous referee 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 .
the work was done in the context of the research network `` accretion onto black holes , compact objects and protostars '' ( tmr grant erb - fmrx - ct98 - 0195 of the european commission ) .
m.revnivtsev acknowledges partial support by rbrf grant 97 - 02 - 16264 and intas grant 933364ext .
arnaud k.a .
1996 , astronomical data analysis software and systems v , eds .
jacoby g. and barnes j. , p17 , asp conf .
series volume 101 .
belloni t. and hasinger g. , 1990 , , 230 , 103 bisnovatyi - kogan , g. s. and blinnikov , s. i. 1976 , soviet astronomy letters , 2 , 191 brandenburg , a. , nordlund , a. , stein , r. f. and torkelsson , u. 1995 , , 446 , 741 brandt , h. , rotschild , r. & swank , j. 1996 , memorie della societa astronomica italiana , 67 , 593 dotani , t. and 9 colleagues 1997 , , 485 , l87 esin a. , mcclintock j.e . , narayan r. , 1997 , , 489 , 865 esin , a. a. , narayan , r. , cui , w. , grove , j. e. and zhang , s. 1998 , , 505 , 854 galeev , a. a. , rosner , r. and vaiana , g. s. 1979 , , 229 , 318 gierliski , m. , zdziarski , a. a. , poutanen , j. , coppi , p. s. , ebisawa , k. and johnson , w. n. 1999 , , 309 , 496 gilfanov m. , churazov e. and revnivtsev m. , 1999 , , 352 , 182 gilfanov m. , churazov e. , revnivtsev m. , 2000 , in preparation haardt , f. and maraschi , l. 1991 , , 380 , l51 hawley , j. f. , gammie , c. f. and balbus , s. a. 1995 , , 440 , 742 liang e.p.t . , price r.h . , 1977 , , 218 , 247 lynden - bell d. and pringle j. e. , 1974 , , 168 , 603 lyubarskii , y. e. , 1997 , , 292 , 679 meyer , f. , liu , b. f. and meyer - hofmeister , e. 2000 , , 354 , l67 miyamoto s. , kitamoto s. , iga s. , hayashida k. and terada k. 1994 , , 435 , 398 muno m. p. , morgan e. h. and remillard r. a. , 1999 , , 527 , 321 poutanen , j. 1998 , theory of black hole accretion disks , eds .
abramowicz , g. bjornsson , j.e .
pringle , cambridge university press , 100 revnivtsev m. , gilfanov m. and churazov e. , 2000 , , submitted , astro - ph9910423 revnivtsev m. , gilfanov m. , churazov e. , 2000 , , submitted , astro - ph/0007092 shakura n. , sunyaev r. , 1973 , , 24 , 337 sunyaev r. , truemper j. , 1979 , , 279 , 506 tanaka y. , shibazaki n. , 1996 , , 34 , 607 thorne k.s . , price r.h . , 1975 , , 195 , l101 tananbaum , h. , gursky , h. , kellogg , e. , giacconi , r. and jones , c. 1972 , , 177 , l5 trudolyubov s. , churazov e. and gilfanov m. , 1999 , astronomy letters , 25 , 718 van der klis m. , 1994 , , 92 , 511 vikhlinin a. , churazov e. and gilfanov m. , 1994 , , 287 , 73 zhang , s. n. , cui , w. , harmon , b. a. , paciesas , w. s. , remillard , r. e. and van paradijs , j. 1997 , , 477 , l95
in this appendix we formally demonstrate the natural fact that in a geometrically thin disk variations of the viscosity at some radius on a time scale shorter than the diffusion time scale causes negligible variations of the mass accretion rate at smaller radii .
consider for simplicity a geometrically thin ( gas pressure dominated ) disk with the constant ratio @xmath61 , where @xmath46 is a half thickness of the disk and @xmath45 is the distance from the compact object .
let us assume that viscosity ( @xmath27 parameter ) suddenly increases in a narrow ring at some distance @xmath62 from the center .
following lyubarskii ( 1997 ) the deviation of the mass accretion rate from the steady state value at the radius @xmath63 is : @xmath64\end{aligned}\ ] ] where time @xmath65 is expressed in units of @xmath66 ( here @xmath67 is a keplerian frequency at the radius @xmath62 ) , @xmath68 , @xmath69 is the bessel function of imaginary argument .
the factor @xmath70 is of the order of @xmath29@xmath71 for the standard geometrically thin ( optically thick ) gas pressure dominated disk and @xmath70 decreases when the thickness of the disk increases . in the limit of @xmath72 ( i.e. at the radii much smaller than the radius @xmath62 where @xmath27 is changing ) the variations of the mass accretion rate
is obviously : @xmath73 the power density spectra associated with the variability in the form ( [ md2 ] ) are shown in fig.[mpds ] . here
the dotted curve corresponds to the case of @xmath74 ( i.e. the typical value in the standard geometrically thin disk ) .
one can see that ( as expected ) virtually no variability is present at high frequencies comparable to @xmath44 . for comparison
the solid line shows the power density spectrum formally calculated for @xmath75 necessarily means that disk is geometrically thick and the above equations , derived in the limit of a geometrically thin disk , are not applicable ] . here
significant variability is present up to high frequencies . | two component x ray spectra ( soft multicolor black body plus harder power law ) are frequently observed from accreting black holes .
these components are presumably associated with the different parts of the accretion flow ( optically thick and optically thin respectively ) in the vicinity of the compact source .
most of the aperiodic variability of the x ray flux on the short time scales is associated with the harder component .
we suggest that drastically different amplitudes of variability of these two components are simply related to the very different viscous time scales in the geometrically thin and geometrically thick parts of the accretion flow .
in the geometrically thin disks variations of viscosity or mass accretion rate occurring at large radius from the black hole on the local dynamical or thermal time scales do not cause any significant variations of the mass accretion rate at smaller radii due to a very long diffusion time .
any variations on the time scales shorter than the diffusion time scale are effectively dampened . on the contrary such variations
can easily survive in the geometrically thick flows and as a result the mass accretion rate in the innermost region of the flow will reflect modulations of the mass accretion rate added to the flow at any distance from the black hole .
therefore if primary instabilities operate on the short time scales then the stability of the soft component ( originating from the geometrically thin and optically thick flow ) and variability of the hard component ( coming from the geometrically thick and optically thin flow ) are naturally explained . for cygnus x-1 overall shape of the power density spectra ( pds ) in the soft and hard spectral states
can be qualitatively explained if the geometrically thin disk is sandwiched by the geometrically thick corona extending in a radial direction up to large distance from the compact object . in the hard state
the thin disk is truncated at some distance from the black hole followed by the geometrically thick flow .
the break in the pds is then associated with the characteristic frequencies in the accretion flow at the thin disk truncation radius .
epsf # 1= # 1#2=.99 = .99 # 1#2#3=.99 = .99 = .99 accretion , accretion disks stars : individual ( cygnus x-1 ) x rays : general [ firstpage ] |
competition between two groups or among a larger number of groups is ubiquitous in business and politics : the decades - long battle between the mac and the pc in the computer industry , between procter & gamble and unilever in the personal products industry , among all major international and local banks in the financial market , and among politicians and interest groups in the world of governance .
all competitors want to increase the number of their supporters and thus increase their chances of success . in gathering supporters , competitors put much effort into persuading skeptics and those opponents who may actually be potential supporters .
this kind of activity is normally modeled as a dynamic process on a complex network in which the nodes are the agents and the links are the interactions between agents .
the goal of these models is to understand how an initially disordered configuration can become an ordered configuration through the interaction between agents . in the context of social science , order means agreement and disorder means disagreement @xcite . most of these models
, the sznajd model @xcite , the voter model @xcite , the majority rule model @xcite , and the social impact model @xcite are based on two - state spin systems which tend to reduce the variability of the initial state and lead to a consensus state in which all the agents share the same opinion .
however this consensus state is not very realistic , since in many real competitions there are always at least two groups that co - exist at the same time .
recently a non - consensus opinion ( nco ) model @xcite was developed , where two opinions @xmath1 and @xmath2 compete and reach a non - consensus stable state . at each time
step each node adopts the opinion of the majority in its `` neighborhood , '' which consists of its nearest neighbors and itself .
when there is a tie , the node does not change its state .
considering also the node s own opinion leads to the non - consensus state .
the dynamics are such that a steady state in which opinions @xmath1 and @xmath2 coexist is quickly reached .
it was conjectured , and supported by intensive simulations @xcite , that the nco model in complex networks belongs to the same universality class as percolation @xcite . here
we test how competition strategies are affected when `` contrarians '' are introduced .
contrarians are agents who hold a strong opinion that is opposition to an opinion held by the rest of the group , and are not influenced by their opinion but can influence them .
we develop a spin - type contrarian opinion ( co ) model in which contrarian agents are introduced into the steady state of the nco model .
the goal of the contrarians is to change the opinions of the current supporters of the rival group @xcite .
we see this strategy in operation , for example , when companies send free new products to potential customers in order to convince them to adopt the product and encourage their friends to do the same .
we can observe it also in political campaigns when candidates `` bribe '' voters by offering favors .
the questions we ask in our model are as follows .
do these free products and bribes work and how ? who are the best individuals to chose as contrarians in order to make most impact on opinion change . in this paper
we introduce , into group @xmath1 , a fraction @xmath0 of contrarians , which are defined to be agents that hold a strong @xmath2 opinion , who will influence those who hold the @xmath1 opinion to change their opinion to @xmath2 .
we study two different strategies of introducing contrarians , ( i ) random choosing of contrarians , and ( ii ) targeted .
we study these strategies on two types of networks , erds - rnyi ( er ) @xcite and scale free ( sf ) @xcite .
we find , for both strategies , that the relative size of the largest cluster in state @xmath1 undergoes a second order phase transition at a critical fraction of contrarians @xmath3 .
moreover we find that , for both networks analyzed here , the targeted strategy is much more efficient than the random strategy .
our results indicate that the observed second order phase transition can be mapped into mean field percolation .
in the nco model@xcite , initially , a fraction @xmath4 of agents with @xmath1 opinions and @xmath5 with @xmath2 opinions are selected at random . at each time step , each network node adopts the majority opinion based on the opinions of its neighbors and itself .
all updates are performed simultaneously at each time step until a steady state is reached in which both opinions coexist , which occurs for @xmath4 above a critical threshold @xmath6 . in our co model ,
the initial state is the final state of the nco model . above @xmath7
we have stable clusters of both @xmath1 or @xmath2 opinions in a network of @xmath8 agents .
we choose , initially , a fraction @xmath0 of @xmath1 opinion agents that are changed into @xmath2 opinion agents and so become contrarians . by contrarian we mean
that , they will never change their opinion but they can influence others .
then we use again the nco dynamics to reach a new steady state . in the new steady state
the agents form again clusters of two opposite opinions above a certain threshold @xmath7 that now depends on @xmath0 . because of the contrarians of type @xmath2 , the @xmath1 clusters become smaller and the @xmath2 clusters increase . in fig .
[ fig : f.12 ] we show a schematic plot of the dynamics of the co model .
we use both random and targeted strategies to choose a fraction @xmath0 of @xmath1 agents that flip into state @xmath2 , and we analyze the results on er and sf networks . in strategy
i we randomly choose a fraction @xmath0 of a agents , and in strategy ii we choose the top @xmath0 percent of the highest degree @xmath1 agents , to become contrarians .
notice that the contrarians act as a quenched disorder in the network @xcite .
we perform simulations of the co model in complex networks , on both er and sf networks . er networks are characterized by a poisson degree distribution with @xmath9 , where @xmath10 is the degree of a node and @xmath11 is the average degree . in sf networks the degree distribution is given by @xmath12 , for @xmath13 , where @xmath14 is the smallest degree , @xmath15 is the degree cutoff and @xmath16 is the exponent characterizing the broadness of the distribution . in all our simulations we use the natural cutoff , which is controlled by @xmath17 @xcite .
we performed all the simulations for @xmath18 network realizations .
we denote by @xmath19 the size of the largest @xmath1 cluster in the steady state .
we study the effect of the contrarians . in fig .
[ fig : f.2 ] we plot @xmath20 as a function of @xmath4 for different values of @xmath0 for er networks under both random and targeted strategies .
the plot shows that there exists a critical value @xmath21 that depends on @xmath0 , below which @xmath22 approaches zero . as @xmath0 increases ,
the largest cluster becomes significantly smaller as well as less robust , as can been seen from the shift of @xmath7 to the larger value . in the inset of fig .
[ fig : f.2 ] , we plot the size of the second largest @xmath1 cluster , @xmath23 , as a function of @xmath4 for different values of @xmath0 .
@xmath23 shows a sharp peak characteristic of a second order phase transition , where @xmath22 is the order parameter and @xmath4 is the control parameter . above a certain value of @xmath24
, the phase transition does not occur because , above @xmath25 , the average connectivity of the @xmath1 nodes decreases dramatically , the networks break into small clusters and no giant component of opinion @xmath1 appears . in fig .
[ fig : f.con ] we show , for both strategies , the average degree @xmath11 of the @xmath1 opinion agents as a function of @xmath4 for different values of @xmath0 . as shown in @xcite for @xmath26
, @xmath27 has a significant increase above @xmath28 where the nodes with opinion @xmath1 are majority .
this is because when these nodes are in the minority group , they have a small average connectivity since the minority group does nt include high degree nodes .
this process is analogous to targeted removing of the high degree nodes . only when they become majority nodes , close to @xmath29 , the original connectivity of the full network
is recovered . however , as @xmath0 increases , we never reach the original average degree of the full network because increasing @xmath0 is equivalent to increasing the quenched disorder .
it is known that for er networks the transition is lost when @xmath30 @xcite . as we can see from the plots , for @xmath31 ( strategy i ) and @xmath32 ( strategy ii )
, @xmath27 drops below @xmath33 , and then the giant component can not be sustained .
the loss of robustness is significantly more pronounced in the targeted strategy compared to the random strategy , as seen in fig .
[ fig : f.2](c ) , where we plot @xmath7 as a function of @xmath0 for both strategies . from this plot
we can see that the targeted strategy is significantly more efficient to annihilate the opinion @xmath1 clusters than the random strategy .
for example , for @xmath34 , the network is 25% less robust in the targeted strategy compared to the random one .
the reason is that the initial state of our model is the final state of the nco model , which above its threshold has clusters of nodes @xmath1 of all sizes .
thus under the random strategy we select nodes at random that are mainly in small @xmath1 clusters . under the targeted strategy
the selection of contrarians from the nodes of highest degree places them mainly in the largest initial @xmath1 cluster where they can influence more than if they were isolated in smaller clusters , as in the random strategy .
the high degree nodes shorten the distances between all the pairs of nodes in a cluster , which allows them to interact more easily .
also , because they have many neighbors , they can influence the opinions of other @xmath1 nodes more efficiently . in order to verify the above arguments , we compute , for our initial condition ( @xmath26 ) before adding the contrarians , the degree distribution of nodes @xmath1 inside and outside the largest cluster . in fig .
[ fig : f.fer](a ) we plot the degree distributions @xmath35 of @xmath1 nodes inside and outside the largest cluster for three different values of @xmath4 above the threshold of the nco model .
notice that the nodes outside the largest cluster have a degree distribution with a high probability of low connectivity .
the probability of low connectivity increases as @xmath4 increases , and thus under a targeted strategy the nodes in those small clusters are almost never designated as contrarians .
thus nodes in the largest component are more likely to be selected under a targeted strategy than under a random one . in order to further test our assumption
, we compute the fraction @xmath36 , defined as the ratio of the number of nodes with degree @xmath10 in the largest @xmath1 cluster and the total number of nodes in all the @xmath1 clusters with the same degree .
when @xmath37 , all the nodes with degree @xmath10 are in the largest @xmath1 cluster . in fig .
[ fig : f.fer](b ) , we plot @xmath36 as a function of @xmath10 for different values of @xmath4 .
as @xmath10 increases , we see that @xmath37 is faster for increasing @xmath4 because the larger @xmath4 is , the larger @xmath19 will be .
thus the highest degree nodes belong to the largest cluster and the lower degree nodes are less likely to be in the largest cluster .
this explains why a targeted strategy is significantly more efficient than a random one . because @xmath0 is our main parameter
, we next investigate the behavior of the system as a function of @xmath0 for different values of @xmath4 . in fig .
[ fig : f.4 ] we plot @xmath22 as a function of @xmath0 for fixed @xmath4 for er networks under both strategies . from the plot we can see that @xmath19 is more robust as @xmath4 is larger , and the behavior of the curve is again characteristic of a second order phase transition
. however this curve seems to be smoother than the transition found above ( see fig . [
fig : f.2 ] ) with @xmath4 as the control parameter . in the inset of fig .
[ fig : f.4 ] ( a ) we plot the first derivative of @xmath22 with respect to @xmath0 for two different system sizes for @xmath38 .
we can see a jump that becomes sharper as the system size increases .
we find the same behavior for other values of @xmath4 above the threshold . in figs .
[ fig : f.6 ] ( a ) and [ fig : f.6 ] ( b ) we show @xmath23 and the first derivative of @xmath22 with respect to @xmath0 for @xmath39 and for different values of @xmath4 .
we find that the peak of @xmath23 and the jump of the derivative of @xmath22 occurs at the same value of @xmath0 .
this behavior is characteristic of a second order phase transition , where the peak of @xmath23 indicates the position of the threshold @xmath3 . in fig .
[ fig : f.6](c ) we plot the critical threshold @xmath3 as a function of @xmath4 for both strategies . comparing the two strategies for the same value of @xmath4 , strategy ii always has the smaller @xmath3 .
this demonstrates again that strategy ii is better because a very small fraction of @xmath0 is enough to destroy the @xmath1 opinion clusters .
next , we present results indicating that the co model is in the same universality class as regular percolation .
percolation in random networks ( _ e.g _ , er)@xcite predicts that at criticality the cluster size distribution of finite clusters @xmath40 with @xmath41 . in fig .
[ fig : f.8 ] we plot @xmath42 for both random and targeted strategies as a function of @xmath43 for finite @xmath1 clusters at criticality . as we can see for both strategies , @xmath44 .
moreover , from @xmath23 we calculate the exponent @xmath45 , which represents how the mean finite size diverges with distance to criticality ( not shown here ) , from which we obtain @xmath46 , as in mean field percolation .
the values of the two exponents we obtain strongly indicate that our co model in er networks is in the same universality class as mean field percolation in complex networks below @xmath25 @xcite .
many real social networks are not er , but instead possess a sf degree distribution .
it is well known that dynamic processes in sf networks propagate significantly more efficiently @xcite than in er networks . for sf networks
we find that the system also undergoes a second order phase transition as in er networks with mean field exponents ( not shown here ) . in fig .
[ fig : f.3 ] ( a ) we plot @xmath7 as a function of @xmath0 for sf networks with @xmath47 . for a certain value of @xmath0 ,
when @xmath48 , we lose the largest @xmath1 cluster .
thus the larger the value of @xmath7 , the less robust the networks . from the plot
, we find that for all values of @xmath0 , strategy ii has much larger @xmath7 than strategy i. this shows that sf networks are significantly less robust under strategy ii than under strategy i , which shows that for sf networks , strategy ii is significantly more efficient compared to strategy i. to further test our conclusion , in fig .
[ fig : f.3 ] ( b ) , we plot @xmath3 as a function of @xmath4 for the same sf networks . as @xmath3 is the minimum concentration of contrarians needed to destroy the largest @xmath1 cluster , thus for the same initial condition , the networks are less robust with smaller @xmath3 than larger one . as shown in fig .
[ fig : f.3](b ) , for the same value of @xmath4 , @xmath3 under strategy ii is always significantly smaller than under strategy i. this result again support our former conclusion that for sf networks , strategy ii is more efficient than strategy i. as mentioned above , this is because the targeted strategy sends most of the contrarians into the largest @xmath1 cluster . in order to test that , in fig . [
fig : f.10 ] we plot @xmath36 as a function of @xmath10 for sf networks . as we can see from fig
. [ fig : f.10 ] , almost all of the high degree nodes ( @xmath49 ) belong to the largest @xmath1 cluster .
this behavior is more pronounced as @xmath4 increases because @xmath19 increases with @xmath4 . if we compare all our results between er and sf networks , we find that both strategies are more efficient for sf networks .
for example , when we compare fig .
[ fig : f.10 ] ( a ) with fig .
[ fig : f.2 ] ( c ) we see that for all the values of @xmath0 , @xmath7 is larger for sf networks than for er networks for both strategies .
the main difference between er and sf networks is that sf networks possess larger hubs than er networks , and thus it is more efficient to destroy the largest @xmath1 cluster .
we also find that the targeted strategy is more efficient in sf than in er networks due to the presence of these large hubs .
for example , when @xmath50 , the sf network is 64% less robust under the targeted strategy than under the random strategy . in er networks , for the same value of @xmath0 , the robustness of the networks decreases only by 17% .
if we compare fig .
[ fig : f.fer ] ( b ) with fig .
[ fig : f.10 ] we see that higher degree nodes are more likely to belong to the largest cluster in sf networks than in er networks , since @xmath51 faster in sf networks compared to er networks .
when two groups compete , either group can use both random and targeted strategies to influence the other group .
will the impact of these strategies differ if the group using them is in the majority , as opposed to being in the minority ? because the largest majority cluster will have a larger average degree than the largest minority cluster , we assume it will be harder to change the opinion of the majority than the minority for @xmath52 . in order to quantitatively understand the effect of contrarians in both a minority group and a majority group , we compute the relative change of the size of the largest minority and majority clusters , @xmath53 , given by @xmath54 where @xmath55 is the size of the largest @xmath1 cluster in our final steady state and @xmath56 is the cluster before adding the contrarians .
notice that @xmath57 ( @xmath58 ) means that the @xmath1 agents are minority ( majority ) . in fig .
[ fig : f.delt ] we plot @xmath53 as a function of @xmath0 for @xmath59 ( minority ) and @xmath60 ( majority ) under both strategies for both er and sf networks . from the plots we can see that below @xmath3 ( marked by arrows in the plots ) , @xmath53 is larger for minority than for majority for the same value of @xmath0 , under both strategies i and ii , and for the two networks used here , er and sf .
thus , as argued above , the minority groups are easier to convince than the majority groups .
moreover , this phenomenon is more pronounced under the targeted strategy than under the random . in the inset of fig .
[ fig : f.delt ] we plot @xmath61 , which is the ratio between @xmath62 of minority and @xmath53 of majority , as a function of @xmath0 .
as we can see for the er network , the contrarians under a targeted strategy are twice more effective when they are in the minority group than when they are in the majority group , while under a random strategy they are approximately 1.5 times more effective .
we can see a similar but more significant tendency in sf networks .
this agrees with empirical fact , where majority groups always have more power than minority groups , and thus it is easier for a majority group to change the opinion of a minority group .
we conclude that our model seems to mimic well the two - group competition in the real world , and that it also reveals some underlying complex phenomena associated with the process .
in introducing contrarians into a system , we have used two strategies : ( i ) random and ( ii ) targeted .
our contrarians hold a strong @xmath2 opinion and the system has two stable opinion @xmath1 clusters and opinion @xmath2 clusters .
we find that , for both strategies , the size of the largest @xmath1 opinion cluster shrinks , as in a phase transition phenomena . as the concentration of contrarians increases , the largest @xmath1 cluster becomes smaller and smaller until it reaches zero at a critical concentration @xmath3 . above @xmath3 ,
the largest @xmath1 cluster disappears .
our results show that the system undergoes a second order phase transition for both control parameters @xmath4 and @xmath0 , behavior that resembles mean field percolation .
we also find that , for both er and sf networks , the targeted strategy is more efficient than the random strategy , because the targeted strategy always sends more contrarians into the largest cluster than the random .
both strategies effect more the minority group and much less the majority group .
note that since sf networks have hubs , both strategies work better in sf networks than in er networks . based on our results ,
we can answer the questions we raise in the introduction .
free products and favors ( `` bribes '' ) do effectively attract more supporters , but the most effective strategy is to target those potential supporters with the most connections and offer the free products and favors to them .
we thank jia shao for discussions , and unmdp , foncyt - pict 2008/0293 , onr , dtra , dfg , eu project epiwork , and the israel science foundation for financial support . in simulations of nco model on 2d lattices , shao et al@xcite suggested that it is in the same universality class as invasion percolation .
we expect that our co model on 2d lattices will be also in the same class as invasion percolation . nodes .
( a ) at @xmath63 , we have a stable state where opinion @xmath1 ( open circle ) and opinion @xmath2 ( filled circle ) coexist .
( b ) at @xmath64 , we change node @xmath33 into a contrarian ( filled square ) , which will hold @xmath2 opinion .
node @xmath65 is now in the local minority opinion while the remaining nodes are not .
notice that node @xmath33 is a contrarian and even if he is in the local minority he does not change his opinion . at the end of this simulation step ,
node @xmath65 is converted into @xmath2 opinion .
( c ) at @xmath66 , node @xmath67 is in the local minority opinion and therefore will be converted into @xmath2 opinion .
( d ) at @xmath68 , the system reaches a stable state where the system breaks into four disconnected clusters , one of them composed of six @xmath2 nodes and the other three with one @xmath1 node .
[ fig : f.12],title="fig:",width=245,height=170 ] nodes .
( a ) at @xmath63 , we have a stable state where opinion @xmath1 ( open circle ) and opinion @xmath2 ( filled circle ) coexist .
( b ) at @xmath64 , we change node @xmath33 into a contrarian ( filled square ) , which will hold @xmath2 opinion .
node @xmath65 is now in the local minority opinion while the remaining nodes are not .
notice that node @xmath33 is a contrarian and even if he is in the local minority he does not change his opinion . at the end of this simulation step ,
node @xmath65 is converted into @xmath2 opinion .
( c ) at @xmath66 , node @xmath67 is in the local minority opinion and therefore will be converted into @xmath2 opinion .
( d ) at @xmath68 , the system reaches a stable state where the system breaks into four disconnected clusters , one of them composed of six @xmath2 nodes and the other three with one @xmath1 node .
[ fig : f.12],title="fig:",width=245,height=170 ] + nodes .
( a ) at @xmath63 , we have a stable state where opinion @xmath1 ( open circle ) and opinion @xmath2 ( filled circle ) coexist .
( b ) at @xmath64 , we change node @xmath33 into a contrarian ( filled square ) , which will hold @xmath2 opinion .
node @xmath65 is now in the local minority opinion while the remaining nodes are not .
notice that node @xmath33 is a contrarian and even if he is in the local minority he does not change his opinion . at the end of this simulation step ,
node @xmath65 is converted into @xmath2 opinion .
( c ) at @xmath66 , node @xmath67 is in the local minority opinion and therefore will be converted into @xmath2 opinion .
( d ) at @xmath68 , the system reaches a stable state where the system breaks into four disconnected clusters , one of them composed of six @xmath2 nodes and the other three with one @xmath1 node .
[ fig : f.12],title="fig:",width=245,height=170 ] nodes .
( a ) at @xmath63 , we have a stable state where opinion @xmath1 ( open circle ) and opinion @xmath2 ( filled circle ) coexist .
( b ) at @xmath64 , we change node @xmath33 into a contrarian ( filled square ) , which will hold @xmath2 opinion .
node @xmath65 is now in the local minority opinion while the remaining nodes are not .
notice that node @xmath33 is a contrarian and even if he is in the local minority he does not change his opinion . at the end of this simulation step ,
node @xmath65 is converted into @xmath2 opinion .
( c ) at @xmath66 , node @xmath67 is in the local minority opinion and therefore will be converted into @xmath2 opinion .
( d ) at @xmath68 , the system reaches a stable state where the system breaks into four disconnected clusters , one of them composed of six @xmath2 nodes and the other three with one @xmath1 node .
[ fig : f.12],title="fig:",width=245,height=170 ] as a function of @xmath4 for different values of @xmath0 for er networks with @xmath69 , and @xmath39 .
( a ) strategy i , @xmath70 ( @xmath71 ) , @xmath72 ( @xmath73 ) @xmath74 ( @xmath75 ) , @xmath76 ( @xmath77 ) , @xmath78 ( @xmath79 ) , @xmath80 ( @xmath81 ) and @xmath82 ( @xmath83 ) .
( b ) strategy ii , @xmath70 ( @xmath71 ) , @xmath72 ( @xmath73 ) , @xmath74 ( @xmath75 ) , @xmath76 ( @xmath77 ) and @xmath84 ( @xmath79 ) . in the inset we plot , using the same symbols as in the main figure , @xmath23 as a function of @xmath4 for both strategies .
( c ) plot of @xmath7 as a function of @xmath0 for strategy i ( @xmath71 ) and ii ( @xmath73).[fig : f.2],title="fig:",width=283,height=245 ] as a function of @xmath4 for different values of @xmath0 for er networks with @xmath69 , and @xmath39 .
( a ) strategy i , @xmath70 ( @xmath71 ) , @xmath72 ( @xmath73 ) @xmath74 ( @xmath75 ) , @xmath76 ( @xmath77 ) , @xmath78 ( @xmath79 ) , @xmath80 ( @xmath81 ) and @xmath82 ( @xmath83 ) . ( b ) strategy ii , @xmath70 ( @xmath71 ) , @xmath72 ( @xmath73 ) , @xmath74 ( @xmath75 ) , @xmath76 ( @xmath77 ) and @xmath84 ( @xmath79 ) . in the inset we plot , using the same symbols as in the main figure , @xmath23 as a function of @xmath4 for both strategies .
( c ) plot of @xmath7 as a function of @xmath0 for strategy i ( @xmath71 ) and ii ( @xmath73).[fig : f.2],title="fig:",width=283,height=245 ] + as a function of @xmath4 for different values of @xmath0 for er networks with @xmath69 , and @xmath39 .
( a ) strategy i , @xmath70 ( @xmath71 ) , @xmath72 ( @xmath73 ) @xmath74 ( @xmath75 ) , @xmath76 ( @xmath77 ) , @xmath78 ( @xmath79 ) , @xmath80 ( @xmath81 ) and @xmath82 ( @xmath83 ) .
( b ) strategy ii , @xmath70 ( @xmath71 ) , @xmath72 ( @xmath73 ) , @xmath74 ( @xmath75 ) , @xmath76 ( @xmath77 ) and @xmath84 ( @xmath79 ) . in the inset
we plot , using the same symbols as in the main figure , @xmath23 as a function of @xmath4 for both strategies .
( c ) plot of @xmath7 as a function of @xmath0 for strategy i ( @xmath71 ) and ii ( @xmath73).[fig : f.2],title="fig:",width=283,height=245 ] as a function of @xmath4 for different value of @xmath0 for er networks with @xmath69 , @xmath39 .
( a ) strategy i , @xmath70 ( @xmath71 ) , @xmath72 ( @xmath73 ) @xmath74 ( @xmath77 ) , @xmath76 ( @xmath79 ) , @xmath78 ( @xmath81 ) , @xmath80 ( @xmath83 ) and @xmath85 ( x ) and ( b ) strategy ii , @xmath70 ( @xmath71 ) , @xmath72 ( @xmath73 ) @xmath74 ( @xmath77 ) , @xmath76 ( @xmath79 ) and @xmath78 ( @xmath81).[fig : f.con],title="fig:",width=340,height=302 ] + as a function of @xmath4 for different value of @xmath0 for er networks with @xmath69 , @xmath39 . ( a ) strategy i , @xmath70 ( @xmath71 ) , @xmath72 ( @xmath73 ) @xmath74 ( @xmath77 ) , @xmath76 ( @xmath79 ) , @xmath78 ( @xmath81 ) , @xmath80 ( @xmath83 ) and @xmath85 ( x ) and ( b ) strategy ii , @xmath70 ( @xmath71 ) , @xmath72 ( @xmath73 ) @xmath74 ( @xmath77 ) , @xmath76 ( @xmath79 ) and @xmath78 ( @xmath81).[fig : f.con],title="fig:",width=340,height=302 ] and @xmath39 , ( a ) degree distribution @xmath35 of @xmath1 nodes as a function of @xmath10 in our initial configuration with different values of @xmath4 , @xmath86 ( full line ) , @xmath38 ( dotted line ) , and @xmath59 ( dashed line ) . in the top , we show @xmath35 as a function of @xmath10 of the finite @xmath1 clusters , and in the bottom the same for the largest @xmath1 cluster .
( b ) plot of @xmath36 as a function of @xmath10 for different value of @xmath4 , @xmath87 ( @xmath71 ) , @xmath78 ( @xmath73 ) and @xmath88 ( @xmath75 ) .
[ fig : f.fer],title="fig:",width=340,height=302 ] + and @xmath39 , ( a ) degree distribution @xmath35 of @xmath1 nodes as a function of @xmath10 in our initial configuration with different values of @xmath4 , @xmath86 ( full line ) , @xmath38 ( dotted line ) , and @xmath59 ( dashed line ) . in the top
, we show @xmath35 as a function of @xmath10 of the finite @xmath1 clusters , and in the bottom the same for the largest @xmath1 cluster .
( b ) plot of @xmath36 as a function of @xmath10 for different value of @xmath4 , @xmath87 ( @xmath71 ) , @xmath78 ( @xmath73 ) and @xmath88 ( @xmath75 ) .
[ fig : f.fer],title="fig:",width=340,height=302 ] as a function of @xmath0 for different values of @xmath4 for er network with @xmath69 and @xmath39 .
( a ) strategy i , @xmath87 ( @xmath71 ) , @xmath78 ( @xmath73 ) @xmath88 ( @xmath75 ) and @xmath80 ( @xmath77 ) .
( b ) strategy ii for @xmath87 ( @xmath71 ) , @xmath78 ( @xmath73 ) and @xmath88 ( @xmath75 ) . in the inset of ( a ) , we plot the first derivative of @xmath89 with respect to @xmath0 ( @xmath90 ) with different system size , @xmath39 ( @xmath71 ) and @xmath91 ( @xmath73).[fig : f.4],title="fig:",width=340,height=302 ] + as a function of @xmath0 for different values of @xmath4 for er network with @xmath69 and @xmath39 .
( a ) strategy i , @xmath87 ( @xmath71 ) , @xmath78 ( @xmath73 ) @xmath88 ( @xmath75 ) and @xmath80 ( @xmath77 ) .
( b ) strategy ii for @xmath87 ( @xmath71 ) , @xmath78 ( @xmath73 ) and @xmath88 ( @xmath75 ) . in the inset of ( a )
, we plot the first derivative of @xmath89 with respect to @xmath0 ( @xmath90 ) with different system size , @xmath39 ( @xmath71 ) and @xmath91 ( @xmath73).[fig : f.4],title="fig:",width=340,height=302 ] as a function of @xmath0 ( top ) and @xmath90 as a function of @xmath0 ( bottom ) for different values of @xmath4 for ( a ) strategy i and ( b ) strategy ii , @xmath87 ( @xmath71 ) , @xmath78 ( @xmath73 ) @xmath88 ( @xmath75 ) for er networks with @xmath92 , @xmath39 .
we can see that in both cases the peak of @xmath23 coincides with the position of the jump , also indicating a second order phase transition . in ( c )
we plot @xmath3 as function of @xmath4 for both strategies , strategy i ( @xmath71 ) and strategy ii ( @xmath73).[fig : f.6],title="fig:",width=283,height=245 ] as a function of @xmath0 ( top ) and @xmath90 as a function of @xmath0 ( bottom ) for different values of @xmath4 for ( a ) strategy i and ( b ) strategy ii , @xmath87 ( @xmath71 ) , @xmath78 ( @xmath73 ) @xmath88 ( @xmath75 ) for er networks with @xmath92 , @xmath39 .
we can see that in both cases the peak of @xmath23 coincides with the position of the jump , also indicating a second order phase transition . in ( c )
we plot @xmath3 as function of @xmath4 for both strategies , strategy i ( @xmath71 ) and strategy ii ( @xmath73).[fig : f.6],title="fig:",width=283,height=245 ] + as a function of @xmath0 ( top ) and @xmath90 as a function of @xmath0 ( bottom ) for different values of @xmath4 for ( a ) strategy i and ( b ) strategy ii , @xmath87 ( @xmath71 ) , @xmath78 ( @xmath73 ) @xmath88 ( @xmath75 ) for er networks with @xmath92 , @xmath39 .
we can see that in both cases the peak of @xmath23 coincides with the position of the jump , also indicating a second order phase transition . in ( c )
we plot @xmath3 as function of @xmath4 for both strategies , strategy i ( @xmath71 ) and strategy ii ( @xmath73).[fig : f.6],title="fig:",width=283,height=245 ] as a function of @xmath43 under strategy i ( @xmath71 ) and ii ( @xmath73 ) for er networks with @xmath92 and @xmath39 at criticality @xmath93 .
the dashed line represents a slope @xmath41 .
this simulation were done over @xmath94 realizations.[fig : f.8],width=529,height=453 ] as a function of @xmath0 and ( b ) @xmath3 as a function of @xmath4 for strategy i ( @xmath71 ) and ii ( @xmath73 ) for sf networks with @xmath95 , @xmath96 and @xmath39.[fig : f.3],title="fig:",width=377,height=340 ] + as a function of @xmath0 and ( b ) @xmath3 as a function of @xmath4 for strategy i ( @xmath71 ) and ii ( @xmath73 ) for sf networks with @xmath95 , @xmath96 and @xmath39.[fig : f.3],title="fig:",width=377,height=340 ] as a function , in linear - log scale , of @xmath10 for er network with @xmath69 and @xmath39 with different value of @xmath4 , @xmath87 ( @xmath71 ) , @xmath78 ( @xmath73 ) @xmath88 ( @xmath75 ) .
the reason of using linear - log scale is that for sf networks @xmath36 increase very fast for small value of @xmath10 [ fig : f.10],width=529,height=453 ] as a function of @xmath0 for @xmath59 ( @xmath71 ) and for @xmath60 ( @xmath73 ) under strategy i ( left plot ) and strategy ii ( right plot ) for ( a ) er networks with @xmath97 , @xmath39 and ( b ) sf networks with @xmath95 , @xmath39 .
the arrows indicate the position of @xmath25 above which there is no phase transition . in the insets we show the ratio @xmath61 between @xmath53 for @xmath59 and @xmath53 for @xmath60 for strategy i ( @xmath98 ) and strategy ii ( x ) .
all the simulation were done for @xmath94 networks realizations.[fig : f.delt],title="fig:",width=377,height=340 ] + as a function of @xmath0 for @xmath59 ( @xmath71 ) and for @xmath60 ( @xmath73 ) under strategy i ( left plot ) and strategy ii ( right plot ) for ( a ) er networks with @xmath97 , @xmath39 and ( b ) sf networks with @xmath95 , @xmath39 .
the arrows indicate the position of @xmath25 above which there is no phase transition . in the insets
we show the ratio @xmath61 between @xmath53 for @xmath59 and @xmath53 for @xmath60 for strategy i ( @xmath98 ) and strategy ii ( x ) .
all the simulation were done for @xmath94 networks realizations.[fig : f.delt],title="fig:",width=377,height=340 ] | we introduce a contrarian opinion ( co ) model in which a fraction @xmath0 of contrarians within a group holds a strong opinion opposite to the opinion held by the rest of the group . at the initial stage ,
stable clusters of two opinions , @xmath1 and @xmath2 exist
. then we introduce contrarians which hold a strong @xmath2 opinion into the opinion @xmath1 group . through their interactions ,
the contrarians are able to decrease the size of the largest @xmath1 opinion cluster , and even destroy it .
we see this kind of method in operation , _ e.g _ when companies send free new products to potential customers in order to convince them to adopt the product and influence others .
we study the co model , using two different strategies , on both erds - rnyi and scale - free networks . in strategy
i , the contrarians are positioned at random . in strategy
ii , the contrarians are chosen to be the highest degrees nodes .
we find that for both strategies the size of the largest @xmath1 cluster decreases to zero as @xmath0 increases as in a phase transition . at a critical threshold value
@xmath3 the system undergoes a second - order phase transition that belongs to the same universality class of mean field percolation .
we find that even for an erds - rnyi type model , where the degrees of the nodes are not so distinct , strategy ii is significantly more effctive in reducing the size of the largest @xmath1 opinion cluster and , at very small values of @xmath0 , the largest @xmath1 opinion cluster is destroyed . |
relativistic jets have been observed in active galactic nuclei ( agns ) and microquasars in our galaxy , and it is believed that they originate from the regions near accreting black holes . to investigate the dynamics of accretion disks and the associated jet formation , we use our newly developed full 3-d grmhd code .
recently , koide , shibata , & kudoh ( 1999 ) have investigated the dynamics of an accretion disk initially threaded by a uniform poloidal magnetic field in a non - rotating corona ( either in a state of steady fall or in hydrostatic equilibrium ) around a non - rotating black hole .
the numerical results show that the disk loses angular momentum by magnetic braking , then falls into the black hole . the infalling motion of the disk , which is faster than in the non - relativistic case because of the general - relativistic effect below @xmath0 ( @xmath1 is the schwarzschild radius ) , is strongly decelerated at the shock formed by the centrifugal force around @xmath2 by the rotation of the disk .
plasmas near the shock are accelerated by the @xmath3 force , which forms bipolar relativistic jets . inside this _ magnetically driven jet _ , the gradient of gas pressure also generates a jet over the shock region ( _ gas - pressure - driven jet _ ) .
this _ two - layered jet structure _ is formed both in a hydrostatic corona and in a steady - state falling corona .
koide et al .
( 2000 ) have also developed a new grmhd code in kerr geometry and have found that , with a rapidly rotating ( @xmath4 ) black - hole magnetosphere , the maximum velocity of the jet is 0.9 c and its terminal velocity 0.85 c. all of the previous 2-d grmhd simulations described here were made assuming axisymmetry with respect to the @xmath5-axis and mirror symmetry with respect to the plane @xmath6 ; the axisymmetric assumption suppressed the azimuthal instabilities .
our basic equations are those of maxwell for the fields and a set of general - relativistic equations representing the plasma , namely the equations of conservation of mass , momentum , and energy for a single - component conducting fluid ( weinberg 1972 ; thorne et al .
in making the simulations , we use these equations with the 3 + 1 formalism ( for details , see koide , shibata , & kudoh 1999 ) .
in order to investigate how accretion disks near black holes evolve under the influence of accretion instabilities such as the magnetorotational instability , the use of a fully 3-d grmhd is essential . in the assumed initial state , the simulation region
is divided into two parts : a background corona around a black hole , and an accretion disk ( fig .
1a and 1b ) .
the coronal plasma is set in a state of transonic free - fall flow , as in the case of the transonic flows with @xmath7 and @xmath8 ; here the sonic point is located at @xmath9 .
the keplerian disk in the corona is set in the following way .
the disk region is located at @xmath10^{1/2}$ ] .
( note that this equation reduces to the newtonian keplerian velocity @xmath11 in the non - relativistic limit @xmath12 ) .
the pressure of both the corona and the disk are assumed equal to that of the transonic solution .
the initial conditions for the entire plasma around the black hole are : @xmath13 @xmath14 @xmath15 where we set @xmath16 ; the smoothing length is @xmath17 .
in addition , there is a magnetic field crossing the accretion disk perpendicularly .
we set it to the wald solution ( wald 1974 ) , which represents the uniform magnetic field around a kerr black hole : @xmath18 , @xmath19 ( where @xmath20 is the lapse function , @xmath21 ) . at the inner edge of the accretion disk ,
the proper alfvn velocity is @xmath22 in a typical case with @xmath23 , where the alfvn velocity in the fiducial observer @xmath24/c^2}}.\ ] ] the plasma beta of the corona at @xmath25 is @xmath26 .
the simulation is performed in the region @xmath27 , @xmath28 , @xmath29 with @xmath30 meshes .
the effective linear mesh widths at @xmath31 and at @xmath32 are @xmath33 and @xmath34 , respectively , while the angular spacings along the polar and azimuthal directions are @xmath35 rad .
a radiative boundary condition is imposed at @xmath36 and at @xmath37 : @xmath38 where the superscripts @xmath39 and @xmath40 denote the time steps and the subscripts 0 and 1 refer to the boundary and to its neighbor meshes , respectively .
the computations were made on an origin 2000 computer with 0.898 gb internal memory , and they used about 47 hours of cpu time for 10000 time steps with @xmath30 meshes .
figure 1 shows the evolution of 3-d simulation performed in the region @xmath41 , @xmath42 , and @xmath43 with @xmath30 meshes .
the parameters used in this simulation are the same as those of the axisymmetric simulations shown in fig .
6 of koide , shibata , & kudoh ( 1999 ) . in this figure
, the colored shading shows the proper mass density ( ( a ) and ( c ) ) and the pressure ( ( b ) and ( d ) ) on logarithmic scales ; the vector plots show the magnetic field ( ( a ) and ( c ) ) and flow velocity ( ( b ) and ( d ) ) .
the black circle represents the black hole .
figures 1a and 1b present the initial conditions , which are the same as in the 2-d simulations ( koide , shibata , & kudoh 1999 ) . at @xmath44 comparing with fig .
6c ( @xmath45 ) in koide et al .
( 1999 ) the jet is less generated . at @xmath46 the jet is generated as 2-d simulation at the earlier time ( @xmath45 ) . at @xmath47 , the jet is clearly created around @xmath48 , which is shown by the enhanced density ( fig .
1c ) and pressure ( fig .
1d ) . as in the 2-d simulations
the jet is generated in a hollowed cylindrical form .
the magnetic field is twisted by the accretion disk and pinched , which increases the magnetic field pressure and generates the jet near the black hole .
the delay of jet formation seems to be due to reduction of shock formation at @xmath2 caused by the additional freedom in the azimuthal direction .
further investigation will be reported elsewhere .
this simulation result is initial and we will perform more simulations and investigate effects of the third dimension .
further results will be reported elsewhere .
is partially supported by nsf atm 9730230 , atm-9870072 , atm-0100997 , and int-9981508 .
the simulations have been performed on origin 2000 at ncsa which is supported by nsf .
koide , s. , shibata , k. , & kudoh , t. 1999 , , 522 , 727 koide , s. , meier , d. l. , shibata , k. , & kudoh , t. 2000 , apj , 536 , 668 thorne , k. s. , price , r. h. , & macdonald , d. a. 1986 , black holes : the membrane paradigm ( new haven : yale univ . press ) wald , r. m. 1974 , , 10 , 1680 weinberg , s. 1972 , gravitation and cosmology ( new york : wiley ) | we have performed a first fully 3-d grmhd simulation with schwarzschild black hole with a free falling corona .
the initial simulation results show that a jet is created as in previous axisymmetric simulations
. however , the time to generate the jet is longer than in the 2-d simulations .
we expect that due to the additional azimuthal dimension the dynamics of jet formation can be modified .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in |
an important part of the application of the boundary element method ( bem ) to physical problems is the calculation of the field , from the solution on the boundary .
this requires the evaluation of integrals which contain ` almost - singular ' integrands which are not properly handled by standard gaussian quadratures .
for example , if we consider a two - dimensional potential problem , such as the laplace or helmholtz equations : @xmath3 where @xmath4 is the wavenumber , the potential @xmath5 at some point in the field will be : @xmath6 where @xmath7 is the green s function for the problem , @xmath8 is the boundary of the domain with normal @xmath9 and variables of integration are indicated by subscript @xmath10 . in the case of the helmholtz equation , the singular behaviour of the green s function will be related to the green s function of the corresponding laplace equation ( * ? ? ? * for example ) .
the green s functions for the laplace equation will have a logarithmic singularity in the planar and axisymmetric case ( where the green s function is proportional to an elliptic integral ) and also in the case of an asymmetric problem in an axisymmetric domain @xcite . the green s function for the two - dimensional laplace equation is @xmath11 . if the boundary integral problem has been solved using the bem , the evaluation of the potential in the field gives rise , on each element , to an integral of the form : @xmath12 where @xmath13 is the local coordinate on the element , @xmath14 is a shape function , typically a polynomial , and @xmath15 and @xmath16 are the field point coordinates in the local coordinate system , figure [ fig : bem : geometry ] .
this integral is non - singular but suffers from an ` offstage singularity ' @xcite if the field point is near the element , so that the argument of the logarithm becomes small . , @xmath17 .
] if it is required to determine the first or second derivatives of the potential , for example , to determine a velocity or velocity gradient in a fluid - dynamical problem , integrals of the form : [ equ : int : deriv ] @xmath18^{1/2 } } + l(t)g(x , y ,
t)\,{\mathrm{d}}t,\\ \label{equ : int:2 } i^{(2)}(x , y ) & = \int_{-1}^{1 } \frac{l(t)}{(x - t)^{2}+y^{2 } } + l(t)h(x , y , t)\,{\mathrm{d}}t , \end{aligned}\ ] ] arise . by analogy with the case
when @xmath2 , we refer to these integrals as ` near - singular ' and ` near - hypersingular ' respectively .
when @xmath2 , the field point lies on the boundary and the integrals must be treated as cauchy principal values @xcite or as a hadamard finite - part @xcite . in this case , a number of procedures exist for the accurate evaluation of the integrals ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* for example ) a recently - developed method @xcite develops quadrature rules which can be used as ` plug - in ' replacements for standard gaussian quadratures in cases where the integrand contains singularities of the form of equations ( [ equ : int ] ) and ( [ equ : int : deriv ] )
. an important feature of these rules is that they require no analysis of the integrand before use .
when the potential is differentiated , yielding equation ( [ equ : int:1 ] ) , the function @xmath19 will usually contain logarithmic singularities which must be properly accounted for , as well as the leading singularity .
in many problems , the green s function will be complicated and the explicit identification of the singularities will be time - consuming .
it is preferable to have a quadrature rule which correctly integrates all of the singularities present without requiring a detailed analysis .
the previously published method @xcite , generates such rules for singular and hypersingular integrals .
this paper extends this method to ` near - singular ' and ` near - hypersingular ' integrals , giving a simple , easily implemented technique producing quadrature rules which are direct replacements for gaussian quadrature in bem codes .
before developing the quadrature rules needed for near - singular integrals , it is worth examining the reason for the breakdown of standard gaussian quadrature . this will also give an indication of when specialized rules should not be used and standard quadratures are better .
consider the integral : @xmath20^{1/2}}\,{\mathrm{d}}t,\\ & = \frac{1}{r}\int_{-1}^{1 } \frac{1}{((t / r)^{2}-2t\cos\theta / r + 1)^{1/2 } } \,{\mathrm{d}}t,\nonumber\end{aligned}\ ] ] where @xmath21 and @xmath22 .
this can be expanded using the generating function for legendre polynomials @xcite : @xmath23 to yield : [ equ : expansion ] @xmath24^{1/2 } } & = \sum_{k=0}^{\infty } \frac{t^{k}}{r^{k+1}}p_{k}(\cos\theta),\quad t < r\\ \label{equ : expansion:2 } & = \sum_{k=0}^{\infty } \frac{r^{k}}{t^{k+1}}p_{k}(\cos\theta),\quad t > r .
\end{aligned}\ ] ] an @xmath25-point gaussian quadrature rule integrates exactly polynomials up to order @xmath26 .
when @xmath27 , the integrand is well approximated by equation ( [ equ : expansion:1 ] ) and the estimate of @xmath28 returned by gauss - legendre quadrature : @xmath29 will be accurate . if , however , @xmath30 , part of the integrand will be given by the inverse power series of equation ( [ equ : expansion:2 ] ) which can not be correctly integrated by the standard gaussian quadrature . likewise , even if @xmath31 but is not large enough to make the terms of equation ( [ equ : expansion:1 ] ) decay fast enough , there will be a large error in the estimate of @xmath28 . looking ahead to the results presented in [ sec : tests ] , it is expected that standard gauss - legendre quadrature will give good results for large @xmath32 and/or in cases where @xmath33 is small .
otherwise , a specialized rule will be necessary .
the algorithm to be developed gives an @xmath25-point rule which integrates a function of the form : @xmath34^{1/2 } } + c(x , y ,
t)\log[(x - t)^{2}+y^{2}]^{1/2 } + d(x , y , t),\ ] ] where @xmath35 , @xmath36 , @xmath37 and @xmath38 are taken to be polynomials of order up to @xmath39 and the integral @xmath40 where @xmath41 are the quadrature points of an @xmath25-point gauss - legendre quadrature and the weights @xmath42 are to be determined .
a previously - developed algorithm @xcite for the computation of quadrature rules gives a method for the evaluation of @xmath42 when @xmath2 .
the approach is conceptually simple the weights are found as the solution to the system of equations : @xmath43w_{j } = m_{i},\quad i=1,\dots,4m,\ ] ] where @xmath44 are the weighted legendre polynomials at the quadrature points @xmath45 : @xmath46^{1/2 } & m+1 \leq i \leq 2m,\\ p_{i-2m-1}(t_{j})/[(x - t_{j})^{2}+y^{2}]^{1/2 } & 2m+1 \leq i \leq 3m,\\ p_{i-3m-1}(t_{j})[(x - t_{j})^{2}+y^{2 } ] & 3m+1 \leq i \leq 4 m .
\end{array } \right.\ ] ] and the moments @xmath47 are the integrals of @xmath48 @xmath49^{1/2}\,{\mathrm{d}}t & m+1 \leq i \leq 2m,\\ \int_{-1}^{1 } p_{i-2m-1}(t)/[(x - t)^{2}+y^{2}]^{1/2}\,{\mathrm{d}}t & 2m+1 \leq i \leq 3m,\\ \int_{-1}^{1 } p_{i-3m-1}(t)[(x - t)^{2}+y^{2}]\,{\mathrm{d}}t & 3m+1 \leq i \leq 4 m .
\end{array } \right.\ ] ] the system of equations is solved using the appropriate lapack solver @xcite , in the least squares sense when @xmath50 and in the minimum norm sense when @xmath51 .
the only issue which must be clarified is the evaluation of the moments , @xmath47 . in the case
when @xmath2 , the integrals are true cauchy principal values or hadamard finite parts and there exist formulae for their evaluation in terms of associated legendre functions @xcite or simple finite part integrals combined with exact gaussian quadratures @xcite . in this case , however , such simple formulae are not available and a different approach is required . the method outlined in [ sec : rules ] for the evaluation of the quadrature weights @xmath42 requires the evaluation of the moments @xmath47 where @xmath52 where weighting function @xmath53 will be one of @xmath54^{1/2}$ ] or @xmath55^{-n/2}$ ] , @xmath56 or 2 .
a general method can be applied to finding the integrals of weighted legendre polynomials , using basic functional relations and simple formulae for the integrals of elementary functions , readily found in standard references @xcite . assuming that integrals of the form @xmath57 can be evaluated ( the formulae required for this paper are given in the appendix ) , the expansion of powers of @xmath13 in terms of legendre polynomials @xcite : [ equ : legendre ] @xmath58 can be used to show that : [ equ : legendre : int ] @xmath59 then
, given the integrals @xmath60 , @xmath61 , the corresponding integrals of the weighted legendre polynomials , @xmath47 , can be evaluated via : @xmath62 and @xmath63 the evaluation can be carried out in place , with values of @xmath47 overwriting @xmath60 . in double precision , the procedure gives accurate results for @xmath64 before overflow errors cause problems . for bem applications , where the shape functions are typically of order no greater than 3 , this causes no special difficulties , but would limit use of the method in other areas .
obviously , for any application other than straight elements with @xmath11 , a rule with @xmath65 will be needed , but it is unlikely that @xmath39 will need to be greater than 32 .
to summarize , the procedure for computing a quadrature rule with @xmath25 points which can integrate functions of the form of equation ( [ equ : quad : func2 ] ) where @xmath35 , @xmath36 , @xmath37 and @xmath38 are polynomials of order up to @xmath39 is as follows : 1 . find the quadrature points @xmath41 for an @xmath25-point gaussian quadrature , using , for example , the method of davis and rabinowitz @xcite ; 2 .
evaluate the weighted legendre polynomials @xmath44 at each of the quadrature points ; 3 .
compute the moments @xmath47 using the method of [ sec : method : integrals ] ; 4 . solve @xmath66w_{j}=m_{i}$ ] using an appropriate solver .
the integral of @xmath1 is then approximated by : @xmath67
the quadrature method developed in [ sec : method ] is tested by applying it to the evaluation of a reference integral . before carrying out these tests ,
it is of some interest to examine the behaviour of the quadrature weights @xmath42 with respect to the field point position .
figure [ fig : deviation ] shows the root mean square difference @xmath68 between the rule of this paper with @xmath69 , @xmath70 and a standard 64-point gaussian quadrature , with @xmath71^{1/2},\ ] ] where @xmath72 are the weights of the standard rule . , @xmath70 .
] the difference between the rules is shown as a function of @xmath32 and @xmath73 .
it is clear that close to the element ( @xmath32 and/or @xmath73 small ) , the weights of the new quadrature are very large , as they have to cancel the large value of the integrand close to the near - singularity .
further from the element , however , the integrand is better approximated by a polynomial and the weights come closer to those of the standard rule .
the test of accuracy is to examine how well the new quadrature rule evaluates a reference integral .
table [ tab : error:2 ] shows the error in computing : @xmath74 for three values of @xmath32 over the range @xmath75 .
the case of @xmath76 was ignored because for odd @xmath9 , @xmath77 which would not allow for a meaningful estimate of relative error which is defined : @xmath78^{1/2},\ ] ] where @xmath79 is the value of @xmath28 estimated by numerical quadrature .
tables [ tab : error:2 ] and [ tab : error:3 ] show the error @xmath80 , incurred using the method of this paper , compared to @xmath81 , the error using standard gauss - legendre quadrature , using low and high order rules .
the calculation is carried out at three values of @xmath32 to examine the change in error with distance from the element and for various values of @xmath9 .
table [ tab : error:2 ] shows the error for @xmath82 , the important range for boundary element calculations , and a rule with @xmath83 and @xmath84 . from the first two columns of errors ,
it is clear that the modified rule is far superior to the standard quadrature : its error is about twelve orders of magnitude lower than that of the gaussian quadrature .
similarly , for @xmath85 , the mean error is much smaller , being no worse than about @xmath86 , rather than @xmath87 .
once the field point is far from the element , however , at @xmath88 , both rules have similar accuracy .
.root - mean - square error in reference integral computed with modified ( @xmath80 ) and standard ( @xmath81 ) quadratures at three values of @xmath32 , @xmath83 , @xmath84 .
[ cols= " < , < , < , < , < , < , < " , ] to examine the error behaviour in more detail , figures [ fig : error : log:164][fig : error:2:164 ] show the relative error in computing : @xmath89^{1/2},\\ i_{0}^{(1)}(x , y ) & = \int_{-1}^{1 } [ ( x - t)^{2}+y^{2}]^{-1/2},\\ i_{0}^{(2)}(x , y ) & = \int_{-1}^{1 } [ ( x - t)^{2}+y^{2}]^{-1},\end{aligned}\ ] ] with error @xmath90 defined as : @xmath91 where @xmath92 is one of @xmath93 , @xmath94 , @xmath95 and @xmath96 is the corresponding estimate using the quadrature rule . in each case ,
@xmath83 and @xmath84 and a sixteen point gaussian quadrature was also applied for comparison . in each case , for large distances from the element , @xmath88 , both quadrature techniques are accurate , with errors of the order of machine precision . when @xmath97 , however , the error incurred using gaussian quadrature is unacceptably large , while the modified rule gives very accurate answers , again of the order of machine precision . as might be expected ,
the error from the gaussian quadrature is smaller as @xmath98 , due to the greater distance from the element , but it is never better than about @xmath99 , insufficient accuracy for most applications . , @xmath84 .
] , @xmath84 . ] , @xmath84 . ]
finally , figure [ fig : error:2:6416 ] illustrates an interesting point about the error behaviour of the quadrature rule as the number of quadrature points is increased .
it shows the error in the near - hypersingular integral @xmath100 evaluated using a rule with @xmath69 and @xmath70 , with the error from a 64-point gaussian quadrature shown for comparison .
the first point is that the gaussian quadrature is able to cope with the singularity for @xmath97 when @xmath101 , because it can integrate polynomials of high enough order to be able to correctly handle the expansion of the integrand in legendre polynomials .
also , as in the previous cases , it can correctly evaluate the integral for @xmath88 . the modified rule , however , has slightly worse error behaviour than for @xmath102 , with the maximum error being higher for @xmath97 and the error at @xmath88 being greater across the full range of @xmath73 .
the error is still small , being less than @xmath86 , but the reason for the increase is unclear .
it appears to be due to an ambiguity in expressing the integrand in terms of legendre polynomials : the near - hypersingular part of the integrand @xmath103 $ ] is a ratio of two polynomials which can be written as a sum of a proper elementary function and a polynomial .
this polynomial term is then represented twice in the quadrature rule , being handled by the unweighted legendre polynomials , @xmath47 , @xmath104 in equation ( [ equ : moments ] ) and by the weighted polynomials .
an interesting question for future developments of the technique will be how best to choose the elementary functions for the quadrature rule to give optimal accuracy for a given @xmath25 . , @xmath70 .
a method of deriving quadrature rules for the evaluation of the ` near - singular ' integrals which arise in the boundary element method has been derived .
the performance of the technique has been assessed by evaluation of reference integrals and it has been found that it outperforms standard gaussian quadrature rules with the same number of nodes for field points close to the element .
the error in the integral increases slightly with the number of points in the rule , a point which is to be investigated in future work .
to evaluate the required integrals of legendre polynomials using the procedure of [ sec : method : integrals ] , we require a method of evaluating the integrals : e. anderson , z. bai , c. bischof , s. blackford , j. demmel , j. dongarra , j. du croz , a. greenbaum , s. hammarling , a. mckenney , and d. sorensen . .
society for industrial and applied mathematics , philadelphia , pa , third edition , 1999 . | a method of deriving quadrature rules has been developed which gives nodes and weights for a gaussian - type rule which integrates functions of the form : @xmath0^{1/2 } } + c(x , y , t)\log[(x - t)^{2}+y^{2}]^{1/2 } + d(x , y , t),\ ] ] without having to explicitly analyze the singularities of @xmath1 or separate it into its components .
the method extends previous work on a similar technique for the evaluation of cauchy principal value or hadamard finite part integrals , in the case when @xmath2 .
the method is tested by evaluating standard reference integrals and its error is found to be comparable to machine precision in the best case . |
is designed to estimate the location and velocity of objects in the surrounding space .
the radar performs sensing by analyzing correlations between sent and received ( analog ) signals . in this note
we describe the digital radar , i.e. we assume that the radar sends and receives finite sequences .
the reduction to digital setting can be carried out in practice , see for example @xcite , and @xcite . throughout this note
we denote by @xmath4 the set of integers @xmath5 equipped with addition and multiplication modulo @xmath0 .
we denote by @xmath6 the vector space of complex valued functions on @xmath7 equipped with the standard inner product @xmath8 , and refer to it as the _ hilbert space of sequences_. we use the notation @xmath9 to denote the unit complex sphere in @xmath10 : @xmath11 we describe the discrete radar model which was derived in @xcite .
we assume that a radar sends a sequence @xmath12 and receives as an echo a sequence @xmath13 .
the relationship between @xmath14 and @xmath15 is given by the following equation : @xmath16 = h(s)[n ] + \mathcal{w}[n ] , \mbox { } n \in \mathbb{z}_n,\ ] ] where @xmath17 , called the _ channel operator _
, is defined by . ]
@xmath18 = \sum_{k=1}^{r } \alpha_k e(\omega_k n ) s[n - \tau_k ] , \mbox { } n \in \mathbb{z}_n,\ ] ] with @xmath19 s the complex - valued attenuation coefficients associated with target @xmath20 , @xmath21 , @xmath22 the time shift associated with target @xmath20 , @xmath23 the frequency shift associated with target @xmath20 , and @xmath24 denotes a _ random _ noise .
the parameter @xmath1 will be called the _
sparsity _ of the channel .
the time - frequency shifts @xmath25 are related to the location and velocity of target @xmath20 .
we denote the plane @xmath26 of all time - frequency shifts by @xmath27 .
we denote by @xmath28 the probability measure on the sample space generated by the random noise and attenuation coefficients .
let us elaborate on the constraint @xmath29 .
in reality @xmath30 . however , we can rescale the received sequence @xmath15 to make @xmath29 .
the rescaling will not change the quality of the detection , as evident from section [ prm ] .
we make the following assumption on the distribution of @xmath24 : _ assumption _ ( * square root cancellation * ) : for every @xmath31 , there exists @xmath32 , such that for any @xmath33 vectors @xmath34 we have @xmath35 note that an additive white gaussian noise ( awgn ) of a constant , i.e. , independent of @xmath0 , signal - to - noise ratio ( snr ) satisfies this assumption .
in addition , we make the following natural assumption on the distribution of the attenuation coefficients @xmath36 of the channel operator : _ assumption _ ( * uniformity * ) : for any measurable subset @xmath37 we have @xmath38 where @xmath39 denotes the unique up to scaling non - negative borel measure on @xmath40 which is invariant under all rotations of @xmath41 , i.e. , elements in @xmath42 .
the last natural assumption that we make is the independence of the noise and of the vector of attenuation coefficients of the channel operator : _ assumption _ ( * independence * ) : the random sequences @xmath43 and @xmath44 are independent .
the main task of the digital radar system is to extract the channel parameters @xmath45 , @xmath46 using @xmath47 and @xmath15 satisfying ( [ discr_channel ] ) .
one of the most popular methods for channel estimation is the pseudo - random ( pr ) method . in this note
we describe the pr method and analyze its performance .
a classical method to estimate the channel parameters in ( [ discr_channel ] ) is the _ pseudo - random method _
it uses two ingredients - the ambiguity function , and a pseudo - random sequence . in order to reduce the noise component in ( [ discr_channel ] ) , it is common to use the ambiguity function that we are going to describe now .
we consider the time - frequency shift operators @xmath48 @xmath49 which act on @xmath50 by@xmath51 [ n]= e(\omega n)\cdot f[n-\tau ] , \mbox { } n \in \mathbb{z}_n \label{ho}\]]the _ ambiguity function _ of two sequences @xmath52 is defined as the @xmath53 matrix @xmath54=\left\langle \pi ( \tau , \omega ) f , g\right\rangle , \text { \ } \tau , \omega \in \mathbb{z } _ { n}. \label{af}\ ] ] [ fc]the restriction of the ambiguity function to a line in the time - frequency plane , can be computed in @xmath55 arithmetic operations using fast fourier transform . for more details , including explicit formulas
see section v of @xcite .
overall , we can compute the entire ambiguity function in @xmath56 operations .
we say that a norm - one sequence @xmath57 is @xmath58-_pseudo - random , _
@xmath59____see figure [ prfigure ] for illustration if for every @xmath60 we have _ _ _ _
@xmath61\right\vert \leq b/\sqrt{n}. \label{pr}\]]there are several constructions of families of pseudo - random ( pr ) sequences in the literature
see @xcite and references therein . for @xmath62 pseudo - random sequence.,height=151 ] consider a pseudo - random sequence @xmath63 , and assume for simplicity that @xmath64 in ( [ pr ] ) .
then we have @xmath65 \label{prm } \\ & = & \left\ { \begin{array}{c } \widetilde{\alpha } _ { k}+\tsum\limits_{j\neq k}\widetilde{\alpha } _
{ j}/\sqrt{n},\text { \ if } \left ( \tau , \omega \right ) = \left ( \tau _ { k},\omega _ { k}\right ) , \text { } 1\leq k\leq r ; \\ \tsum\limits_{j}\widehat{\alpha } _ { j}/\sqrt{n},\text { \ \ \ \ \ otherwise , \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ } \end{array}\right .
\notag\end{aligned}\]]where @xmath66 @xmath67 @xmath68 are certain multiples of the @xmath69 s by complex numbers of absolute value less or equal to one . in particular , we can compute the time - frequency parameter @xmath70 if the associated attenuation coefficient @xmath71 is sufficiently large , i.e. , it appears as a peak " of @xmath72 let @xmath3 .
we say that at @xmath73 the ambiguity function of @xmath74 and @xmath75 has @xmath76-peak , if @xmath77 \right| \geq n^{-1/2 + \delta}.\ ] ] below we describe see figure [ three - peaks]the pr method . for a pseudo - random sequence @xmath47 , and a channel of sparsity three.,height=151 ] * input : * : : @xmath78 pseudo - random sequence @xmath12 , the echo @xmath15 , and a parameter @xmath3
. * output : * : : @xmath79 channel parameters .
compute @xmath80 on @xmath27 and return those time - frequency shifts at which the @xmath76-peaks occur .
we call the above computational scheme the * pr method with parameter @xmath76*. notice that the arithmetic complexity of the pr method is @xmath81 using remark [ fc ] .
first , we introduce two important quantities that measure the performance of a detection scheme .
these are the probability of detection and the expected number of false targets .
then , we formulate the main result of this note which provides a quantitative statement about the performance of the pr method .
assume that we have @xmath1 targets out of the @xmath33 possible see section model .
also , assume that some random data is associated with these targets and influencing the performance of a detection scheme .
for example , in our setting the random data consists of ( i ) the attenuation coefficients associated with the targets , and ( ii ) the noise . in general
, we model the randomness of the data by associating a probability space @xmath82 to the set of targets . for every @xmath83 in the space @xmath84
we denote by @xmath85 and @xmath86 , the number of true and false targets detected by the scheme , respectively .
we define the _ probability of detection _ by @xmath87 and the _ expected number of false targets _ by @xmath88 the main result of this note is the following : [ main_thm ] assume the channel operator ( [ operator ] ) satisfies the uniformity , square root cancellation , and independence assumptions . then for @xmath89 , where @xmath3 , we have @xmath90 , and @xmath91 , as @xmath92 , for the pr method with parameter @xmath93 .
we suspect that the true rate of convergence in @xmath90 , and @xmath94 , as @xmath92 , is polynomial .
in fact , our proof of theorem [ main_thm ] confirms that the rate of convergence is at least polynomial .
the obtained estimates in theorem [ main_thm ] show that the pr method is effective in terms of performance of detection , in the regime @xmath2 , for @xmath3 .
we would like to note that if @xmath95 , for @xmath96 , then the performance of the pr method deteriorates since the noise influence becomes dominant .
before giving a formal proof , we provide a sketch .
to detect the parameters of the channel operator @xmath17 , the pr method evaluates the ambiguity function of @xmath15 , and @xmath47 at @xmath97 : @xmath98 let us denote by @xmath99 the @xmath20-th cross term , and by @xmath100 the @xmath20-th noise component .
then we have @xmath101 the parameter @xmath97 is detectable by the pr method if the main term @xmath19 is much larger than the @xmath20-th cross term @xmath102 , and the noise component @xmath103 . for a random point on the unit sphere @xmath9 , by concentration "
most of its coordinates are of absolute value approximately equal to @xmath104 .
thus , if @xmath105 , the magnitude of most of @xmath19 s is greater than @xmath106 .
another instance of the concentration phenomenon guarantees that for most channels , the magnitude of the cross term @xmath102 is smaller than @xmath106 .
finally , the square root cancellation assumption on the noise guarantees that the magnitude of the noise term @xmath103 is much smaller than @xmath106 .
we begin the formal proof of the theorem with auxiliary lemmata . in section [ proofs_section ]
we prove these statements .
[ large_coord ] let @xmath36 be a uniformly chosen point on @xmath9 , and fix @xmath107 .
there exists @xmath108 ( independent of @xmath1 and @xmath20 ) , such that for every @xmath109 we have @xmath110 [ prob_lem ] let @xmath111 be events in a probability space @xmath82 , such that @xmath112 , @xmath113 , for some @xmath3 .
then for the event @xmath114 where @xmath115 we have @xmath116 [ mixed_terms ] let @xmath117 , and let @xmath118 , @xmath119 , be vectors satisfying @xmath120 then for any @xmath3 , there exists @xmath121 , such that for a uniformly chosen random point @xmath122 we have @xmath123 assume that @xmath2 , for some @xmath3 .
let @xmath124 , where @xmath47 is a ( @xmath64)-pseudo - random sequence in @xmath125 , @xmath17 is a channel of sparsity @xmath1 with uniformly distributed attenuation coefficients given by ( [ operator ] ) , and @xmath24 satisfies the square root cancellation assumption .
we denote @xmath126 , and assume that at the receiver we perform pr method with parameter @xmath93 . *
( a ) * proof of @xmath90 as @xmath92 " .
we consider two cases . * case 1 .
* @xmath127 .
denote by @xmath128 for @xmath113 . since @xmath2 by lemma [ large_coord ] there exists @xmath129 such that
we have @xmath130 therefore , for sufficiently large @xmath0 , we have @xmath131 denote by @xmath132 by lemma [ prob_lem ] we have @xmath133 since @xmath2 , we have @xmath134
. therefore , by lemma [ mixed_terms ] , there exists @xmath121 such that @xmath135 @xmath136 it follows from ( [ prm ] ) , ( [ eq_1 ] ) , ( [ eq_2 ] ) , the square root cancellation and independence assumptions , that with probability greater or equal than @xmath137 , at least @xmath138 of the channel parameters of @xmath17 are detectable .
the latter implies that @xmath139 for @xmath0 sufficiently large .
therefore we have @xmath90 as @xmath92 .
* case 2 . * @xmath140 .
by lemma [ large_coord ] , there exists @xmath129 such that @xmath141 by cauchy - schwartz inequality we have for all @xmath113 : @xmath142 it follows from ( [ prm ] ) , square root cancellation assumption on the noise , and the last two inequalities that for sufficiently large @xmath0 , all @xmath1 channel parameters of the operator @xmath17 are detectable with probability greater or equal than @xmath143 . therefore ,
for sufficiently large @xmath0 we have @xmath144 the latter implies that @xmath90 as @xmath92 . *
( b ) * proof of @xmath94 as @xmath92 " .
* @xmath145 . by lemma [ mixed_terms ]
, there exists @xmath121 such that we have @xmath146 @xmath147 where @xmath148 is the set of all channel parameters of @xmath17 .
it follows from ( [ prm ] ) , the square root cancellation and independence assumptions , and the last inequality that with probability greater or equal than @xmath149 the pr method will not detect any wrong channel parameters .
the latter implies that @xmath94 as @xmath92 .
* case 2 . * @xmath150 .
by cauchy - shwartz inequality we have that for every @xmath151 : @xmath152 it follows from ( [ prm ] ) , the square root cancellation assumption on the noise , and the last inequality that there exists @xmath32 such that for @xmath0 sufficiently large we have @xmath153 the latter implies that the pr method with parameter @xmath93 satisfies @xmath94 as @xmath92 .
we identify the borel probability space on @xmath9 invariant under all rotations with the borel probability space @xmath154 of the real unit sphere invariant under all rotations . recall that @xmath155 without loss of generality , it is enough to prove the statement for @xmath156 .
let @xmath157 .
we use the notation @xmath82 for the probability space on @xmath154 invariant under all rotations in @xmath158 . for any @xmath159 , and any dimension @xmath160
, we denote the real sphere of radius @xmath161 in @xmath162 by @xmath163 : @xmath164 let @xmath96 . by fubini s theorem and using the homogeneity of the lebesgue measure we get @xmath165 @xmath166 @xmath167
it is well known that @xmath168 therefore , for @xmath1 large enough we have @xmath169 @xmath170 finally , the containment @xmath171 together with last inequality imply the statement of the lemma .
let @xmath96 .
we proceed similarly to the proof of lemma [ large_coord ] .
since @xmath179 @xmath180 and @xmath181 we conclude that there exists @xmath129 such that @xmath182 if @xmath183 , then there exists @xmath184 such that @xmath185 this implies that there exists @xmath186 such that @xmath187 by the rotation invariance of the lebesgue measure on @xmath154 , it follows that there exists @xmath121 such that for any set of directions @xmath188 we have @xmath189 @xmath190 the latter implies the statement of the lemma .
* acknowledgements .
* we are grateful to our collaborators a. sayeed , k. scheim and o. schwartz , for many discussions related to the research reported in these notes . also , we thank anonymous referees for numerous suggestions . and , finally , we are grateful to g. dasarathy who kindly agreed to present this paper at the conference . 99 fish a. , gurevich s. , hadani r. , sayeed a. , and schwartz o. , delay - doppler channel estimation with almost linear complexity
. _ ieee transactions on information theory , volume 59 , issue 11 , 7632 - 7644 , 2013 .
_ gurevich s. , hadani r. , and sochen n. , the finite harmonic oscillator and its applications to sequences , communication and radar .
_ ieee transactions on information theory , vol .
9 , september 2008 . _ | a performance of the pseudo - random method for the radar detection is analyzed .
the radar sends a pseudo - random sequence of length @xmath0 , and receives echo from @xmath1 targets .
we assume the natural assumptions of uniformity on the channel and of the square root cancellation on the noise . then for @xmath2 , where @xmath3 , the following holds : ( i ) the probability of detection goes to one , and ( ii ) the expected number of false targets goes to zero , as @xmath0 goes to infinity . |
in the case of lmxbs containing neutron or quark stars we propose a new model of twin peak hf qpos assuming switching of the twin oscillatory modes creating sequences of the lower and upper hf qpos at a resonant point .
according to such a resonant switch ( rs ) model non - linear resonant phenomena will cause excitation of a new oscillatory mode ( or two new oscillatory modes ) and vanishing of one of the previously acting modes ( or both the previous modes ) , i.e. , switching from one pair of the oscillatory modes to other pair of them that will be acting up to the following resonant point @xcite .
we assume two resonant points at the disc radii @xmath6 and @xmath7 , with observed frequencies @xmath8 , @xmath9 and @xmath10 , @xmath11 , being in commensurable ratios @xmath12 and @xmath13 .
these resonant frequencies are determined by the energy switch effect ( ( * ? ? ?
* trk 2009 ) ) ; observations put the restrictions @xmath14 and @xmath15 . in the region covering the resonant point at @xmath6
the twin oscillatory modes with the upper ( lower ) frequency are determined by the function @xmath16 ( @xmath17 ) . near the inner resonant point at @xmath7 different oscillatory modes given by the frequency functions @xmath18 and
@xmath19 occur ( @xmath20 is the dimensionless radius ) .
we assume all the frequency functions to be determined by combinations of the orbital and epicyclic frequencies of the geodesic motion in the kerr backgrounds .
such a simplification is correct with high precision for neutron ( quark ) stars with large masses , close to the maximum mass allowed for a given equation of state .
the frequency functions have to meet the observationally given resonant frequencies . in the framework of the simple rs model , when two resonant points and two pairs of the frequency functions are assumed , this requirement enables determination of the neutron ( quark ) star parameters .
the `` shooting '' of the frequency functions to the resonant points can be realized efficiently in two steps .
independence of the frequency ratio on the mass parameter @xmath3 implies the conditions @xmath21 giving relations for the spin @xmath2 in terms of the dimensionless radius @xmath22 and the resonant frequency ratio @xmath23 .
they can be expressed in the form @xmath24 and @xmath25 , or in an inverse form @xmath26 and @xmath27 . at the resonant radii
the conditions @xmath28 are satisfied along the functions @xmath29 and @xmath30 which are obtained by using the functions @xmath24 and @xmath25 .
the parameters of the neutron ( quark ) star are then given by the condition @xmath31 the condition ( [ rs ] ) determines @xmath3 and @xmath2 precisely , if the resonant frequencies are determined precisely .
if an error occurs in determination of the resonant frequencies , as naturally expected , our method gives related intervals of acceptable values of mass and spin parameter of the neutron ( quark ) star .
we test the rs model in the case of the atoll 4u 1636@xmath453 source that seems to be the best possibility due to the character of the observational data demonstrating clearly existence of two `` resonant points '' where the energy switch effect occurs . using the results of @xcite ,
the resonant frequencies determined by the energy switch effect are given in the outer resonant point with frequency ratio @xmath32 by the frequency intervals @xmath33 and at the inner resonant point with frequency ratio @xmath34 there is @xmath35 we consider the standard specific models of the twin oscillations based on the orbital and epicyclic geodetical frequencies and in the framework of the rs model we determine range of allowed values of spin @xmath2 and mass @xmath3 of the neutron star located at the atoll source 4u 1636@xmath453 .
the predicted ranges of the neutron star parameters are strongly dependent on the twin modes applied in the rs model .
the mass and spin estimates of the rs model have to be confronted with restrictions on the neutron star mass and spin implied by the theoretical models of neutron ( quark ) stars and the observed rotational frequency of the neutron star at 4u 1636@xmath453 .
we consider as acceptable upper values @xmath36 , @xmath37 and demonstrate that for some of the oscillatory modes used in the rs model the predicted parameters of the neutron star are unacceptable , being too high . among acceptable rs models the most promising are those combining the relativistic precession ( where @xmath38 , @xmath39 ) and the total precession ( where @xmath40 , @xmath41 ) frequency relations or their modifications . the switch of the oscillating modes is not given by the resonant phenomena necessarily .
the cause could be , e.g. , given by the magnetic field , then the alfvn wave model ( @xcite ) can be relevant
. + + * acknowledgements . *
we thank grant gar 202/09/0772 and the project cz.1.07/2.3.00/20.0071 `` synergy '' supporting international collaboration of the institute of physics at su opava . | we present a resonant switch ( rs ) model of twin peak high - frequency quasi - periodic oscillations ( hf qpos ) , assuming switch of twin oscillations at a resonant point , where frequencies of the upper and lower oscillations @xmath0 and @xmath1 become to be commensurable and the twin oscillations change from one pair of the oscillating modes ( corresponding to a specific model of hf qpos ) to some other pair due to non - linear resonant phenomena .
the rs model enables to determine range of allowed values of spin @xmath2 and mass @xmath3 of the neutron star located at the atoll source 4u 1636@xmath453 where two resonant points are observed at frequency ratios @xmath5 , 5:4 . |
the central region of milky way is a potential site for the production of effective particle acceleration and copious @xmath0-ray emission .
tev @xmath0-ray emission from the galactic center(gc ) had been reported by 4 independent groups in recent years : cangaroo ( @xcite ) , whipple ( @xcite ) , hess(@xcite ) , and magic ( @xcite ) .
one possibility for this tev @xmath0-ray emission source is in the whole diffuse 10 pc region , which is proposed to be related to the massive black hole sgr a@xmath1 harbored in our galactic center ( @xcite ) .
there are several radiation mechanisms for the production of tev @xmath0-ray emission .
one of these mechanisms proposed that the tev @xmath0-rays can be produced indirectly through the processes of @xmath2@xmath3-decay when relativistic protons are injected into and interact with the interstellar medium ( ism ) , called hadronic model ( @xcite ) .
if this is the case , by assuming that the initial injected protons spectrum follows a power - law plus a high energy cut - off , we can predict the spectrum of the injected relativistic protons by comparing the spectral energy distribution ( sed ) of the tev @xmath0-ray emission produced by hadronic model with observation data from hess .
to calculate the spectrum of the injected relativistic protons , we consider that the tev @xmath0-ray detected by hess is produced through the hadronic model , and the injection of the relativistic protons required by the hadronic model came from the jet of the black hole sgr a@xmath1 when it captures and tidally disrupts a star(@xcite ) . the tidal disruption of stars by massive black hole sgr a@xmath1 refers to main sequence ( ms ) stars and red giants .
for these two cases , the total energy carried by the jet is exactly the same as 2.7@xmath410@xmath5erg , which is believed to be enough for the production of the injected relativistic protons required by the hadronic model .
the definite difference resulting from the two cases are the diffusion timescale of the injected protons : 800 yr for the case of mss and 2.11@xmath410@xmath6yr for the red giants , respectively ( @xcite ) .
three kinds of diffusion mechanisms are involved for the propagation of the relativistic protons when they are injected into and interact with the ism : ( 1 ) the effective confinement of protons ( ecp ) , ( 2 ) the kolmogorov - type turbulence ( ktt ) , ( 3 ) the bohm diffusion ( bd ) .
for these three mechanisms , the diffusion coefficient depends on the proton energy , given by the formula of @xmath7=10@xmath8(@xmath9/1@xmath10)@xmath11@xmath12 @xmath13 .
the case of ecp corresponds to @xmath14=0.5 and@xmath12=10@xmath15 , ktt corresponds to @xmath14=0.3 and@xmath12=0.15 , and bd corresponds to @xmath14=1.0 and@xmath12=10@xmath16 . giving the initial injected relativistic protons spectrum follows a power - law plus a high energy cut - off , at a given time of @xmath17 and the source distance of @xmath18 , the sed of the tev @xmath0-ray emission produced through the model discussed above
can be calculated through @xmath19=@xmath20@xmath21@xmath22/@xmath23 , where @xmath9@xmath24 is the energy of the emitted @xmath0-ray photons , @xmath21 is the total volume of the ism , @xmath25@xmath24 is the emissivity of @xmath0-ray photons , and @xmath23=4@xmath2(@xmath26@xmath27)@xmath28 , where @xmath26@xmath27 is the distance from the observation to the source , and we adopt @xmath26@xmath27=8@xmath29 hereafter .
we address the sed of the emitted @xmath0-rays in the following cases : ( 1 ) the different power - law index @xmath30 of injected protons with values of 1.7 , 1.8 , 1.9 , 2.0 , 2.1 , 2.2 ; ( 2)the mss and red giants are tidally disrupted by massive black hole sgr @xmath31 ; ( 3)three diffusion mechanisms for the cases of ecp , ktt , and bd . by comparing the theoretical spectrum with the observation data from hess within the central 10pc of our galaxy , we find that the theoretical energy distribution can fit best with the observation data when the type of star disrupted by black hole is a red giant and the diffusion mechanism of protons is the effective confinement of protons ( ecp ) .
in such a case , we derived that the power - law index of the injected relativistic protons is -1.9 .
we have investigated the spectrum of the injected relativistic protons required by the hadronic model to produce the 10 tev @xmath0-ray emission detected by hess in the center of our galaxy . comparing with the observations , we find that the power - law index of the initial spectrum of the injected protons should be -1.9 .
this result is based on that the tidal disruption of a red giant by the black hole sgr @xmath31 is considered and the propagation of protons in the target gas is the ecp scenario . | about 10tev @xmath0-ray emission within 10pc region from the galactic center had been reported by 4 independent groups . considering that this tev @xmath0-ray emission is produced via a hadronic model , and the relativistic protons came from the tidal disruption of stars by massive black holes , we investigate the spectral nature of the injected relativistic protons required by the hadronic model .
the calculation was carried on the tidal disruption of the different types of stars and the different propagation mechanisms of protons in the interstellar medium .
compared with the observation data from hess , we find for the best fitting that the power - law index of the spectrum of the injected protons is about -1.9 , when a red giant star is tidally disrupted , and the effective confinement of protons diffusion mechanism is adopted . |
the quantized universal enveloping algebra @xmath7 associated with a symmetrizable kac
moody lie algebra @xmath8 was introduced independently by drinfeld @xcite and jimbo @xcite in their study of two dimensional solvable lattice models in statistical mechanics .
the parameter @xmath9 corresponds to the temperature of the underlying model .
kashiwara @xcite showed that at zero temperature or @xmath10 the representations of @xmath7 have bases , which he coined crystal bases , with a beautiful combinatorial structure and favorable properties such as uniqueness and stability under tensor products .
the underlying algebras @xmath8 of affine crystals are affine kac
moody algebras .
there are two main categories of affine crystals : ( 1 ) crystal bases of infinite dimensional integrable highest weight @xmath7-modules , and ( 2 ) crystal bases of finite dimensional @xmath11-modules , where @xmath11 is a subalgebra of @xmath7 without the degree operator .
the former category is well - understood .
for instance it is known @xcite that an irreducible integrable @xmath7-module has a unique crystal basis . on the other hand ,
the latter is still far from well - understood .
it is not even known which finite dimensional @xmath11-modules have a crystal basis . in @xcite
it was conjectured that there exists a family of finite dimensional @xmath11-modules @xmath12 having crystal bases @xmath13 ( @xmath14 ) , where @xmath15 is the number of vertices in the dynkin diagram of @xmath8 .
if such crystals indeed exist , one can define one dimensional configuration sums , which play an important role in the study of phase transitions in two dimensional exactly solvable lattice models .
if @xmath8 is of type @xmath16 , the existence of the crystal @xmath17 is settled in @xcite , and the one dimensional configuration sums contain the kostka polynomials @xcite , which turn up in the theory of symmetric functions @xcite , combinatorics @xcite , the study of subgroups of finite abelian groups @xcite , and kazhdan lusztig theory @xcite . in certain limits
they are branching functions of integrable highest weight modules @xcite .
for all affine types and @xmath18 it is shown in @xcite that @xmath19 exists ( but not that it has the desired decomposition as a classical crystal in all cases ) . for type @xmath1
, the structure of @xmath19 has been described explicitly in @xcite .
the existence and structure of @xmath17 has been proven for various other subcases @xcite .
classical finite dimensional crystals of type @xmath4 and @xmath3 can be embedded into crystals of type @xmath5 @xcite .
this embedding has been extended to certain perfect affine crystals @xcite .
following this idea we consider `` virtual '' crystals @xmath20 for types @xmath21 , @xmath1 , @xmath22 ( the dynkin diagram @xmath1 with different distinguished vertex @xmath23 ) , and @xmath2 by naturally extending these embeddings to finite affine crystals .
we conjecture that these virtual crystals @xmath20 are exactly the conjectured crystals @xmath17 of @xcite .
this is proven for @xmath18 and types @xmath1 , @xmath22 , and @xmath2 . for type @xmath22
the work was already done by combining @xcite and @xcite . as further evidence
we show that the virtual crystals have the expected decomposition into classical components for all @xmath24 in type @xmath21 . in all cases
the virtual crystal @xmath20 is shown to have at least the desired classical components .
we give an explicit characterization of elements in @xmath25 , which is used to establish the equality between the one - dimensional sums ( defined in terms of the virtual crystals @xmath25 ) and the fermionic formulas mentioned below .
crystal theory provides a general setting for the definition of one dimensional configuration sums , denoted by @xmath26 .
explicit formulas @xmath27 , called fermionic formulas , for @xmath26 have recently been conjectured in @xcite .
fermionic formulas are expressions for certain polynomials in @xmath9 ( or @xmath9-series ) which are sums of products of @xmath9-binomial coefficients .
these expressions play a preponderant role in bethe ansatz studies of spin chain models . at @xmath28
they give the number of solutions to bethe equations of the underlying model .
kirillov and reshetikhin @xcite obtained fermionic expressions for the kostka polynomials from the bethe ansatz study of the @xmath29 model . in a separate development kedem and
mccoy @xcite discovered fermionic expressions for the branching functions of the coset pair @xmath30 based on the bethe ansatz study of the 3-state potts chain .
this led to a flurry of works on fermionic expressions for characters and branching functions of conformal field theories ( see @xcite for references ) . at present many of these
fermionic expressions are understood as certain limits of @xmath27 or its level truncation .
branching functions and one dimensional configuration sums are , however , not the only places where fermionic formulas have appeared .
kirillov and reshetikhin @xcite conjectured that for @xmath8 of nontwisted affine type , the coefficients of the decomposition of the representations of @xmath11 naturally associated with multiples of the fundamental weights into direct sums of irreducible representations of @xmath31 are given by the fermionic formulas at @xmath28 . here
@xmath32 is the simple lie algebra associated to the kac moody algebra @xmath8 .
chari @xcite proved a number of cases of this conjecture .
recently , it was conjectured by lusztig @xcite and nakajima @xcite that quiver varieties are also related to fermionic formulas . the conjecture @xmath33 was proven in @xcite for type @xmath34 .
the main technique used in the proof is a statistic preserving bijection between crystals and rigged configurations .
rigged configurations are combinatorial objects which label all terms in the fermionic formula and also originate from the bethe ansatz @xcite . here
we prove the conjecture @xmath33 of @xcite for types @xmath21 , @xmath1 and @xmath2 for tensor products of crystals of the form @xmath19 , assuming that @xmath35 for type @xmath21 .
we also conjecture a new fermionic formula of type @xmath22 .
the proof of @xmath33 is achieved by an explicit characterization of the rigged configurations in the image of the embedding of the above crystals into crystals of type @xmath36 combined with the bijection between type @xmath36 crystals and rigged configurations @xcite .
an important tool that is used in the proofs and the characterizations is the contragredient duality on crystals and rigged configurations .
section [ sec : crystal ] gives a review of crystal theory , in particular the definition of crystal graphs , dual crystals , simple crystals , the combinatorial @xmath37-matrix , the ( intrinsic ) energy function and the definition of the one dimensional configuration sums . in section [ sec : type a ] the combinatorial structure of crystals of type @xmath34 is discussed in detail , and all results and definitions of section [ sec : crystal ] are stated in terms of operations on young tableaux .
irreducible components of tensor products of crystals of type @xmath34 can be parametrized in terms of littlewood
richardson tableaux .
these are introduced in section [ sec : lr ] .
they are of importance in section [ sec : rc ] where we define rigged configurations and recall the bijection between littlewood
richardson tableaux and rigged configurations of @xcite .
the relation between the duality and this bijection is given in theorem [ cdual rc ] .
virtual crystals @xmath38 for types @xmath21 , @xmath1 , @xmath22 , and @xmath2 are introduced in section [ sec : three types ] .
their explicit characterization for @xmath18 is given in section [ subsec : char ] in terms of classical virtual crystals @xmath39 .
finally , in section [ sec : fermi ] the fermionic formulas for @xmath18 and types @xmath21 , @xmath1 and @xmath2 are proved .
the proof relies on the main theorem [ thm rc prop ] which gives a characterization of the rigged configurations in the image of the virtual crystals under the bijection of section [ sec : rc ] . for type @xmath22 a similar characterization
is given in conjecture [ conj : rc atd ] which yields a new fermionic formula given in section [ sec : fermi atd ] .
appendix [ app proof ] is reserved for the proof of theorem [ thm rc prop ] .
finally , it worth mentioning that the techniques of this paper have been applied in @xcite to obtain fermionic formulas for level - restricted one dimensional configuration sums of type @xmath2 . the self - duality property of rigged configurations of type @xmath40 as stated in theorem [ thm rc prop ] was first observed by tim baker .
the authors thank him for this information .
in this section some results from affine crystal theory are reviewed .
in addition we show that the tensor category of simple crystal bases of irreducible finite - dimensional modules over quantized affine algebras , can be enhanced by including a grading function which we call the intrinsic energy .
this result ( proposition [ pp : graded tensor ] ) is central to the definition of the energy function of a virtual crystal , notions that are introduced in section [ sec : three types ] .
a prerequisite for the explicit formula for the intrinsic energy of a tensor product , is the discussion in section [ subsec : iew ] of isomorphisms and energy functions indexed by reduced words .
let @xmath8 be a symmetrizable kac - moody algebra , @xmath41 the weight lattice , @xmath42 the index set for the vertices of the dynkin diagram of @xmath8 , @xmath43 the simple roots , and @xmath44 the simple coroots .
let @xmath7 be the quantized universal enveloping algebra of @xmath8 @xcite .
a @xmath7-crystal is a nonempty set @xmath45 equipped with maps @xmath46 and @xmath47 for all @xmath48 , satisfying @xmath49 here for @xmath50 @xmath51 ( it is assumed that @xmath52 for all @xmath48 and @xmath53 . ) a @xmath7-crystal @xmath45 can be viewed as a directed edge - colored graph ( the crystal graph ) whose vertices are the elements of @xmath45 , with a directed edge from @xmath54 to @xmath55 labeled @xmath48 , if and only if @xmath56 .
a morphism @xmath57 of @xmath7-crystals is a map @xmath58 such that @xmath59 let @xmath60 be @xmath7-crystals .
the cartesian product @xmath61 has the structure of a @xmath7-crystal using the so - called signature rule @xcite .
the resulting crystal is denoted @xmath62 and its elements @xmath63 are written @xmath64 where @xmath65 . the reader is warned that our convention is opposite to that of kashiwara @xcite .
fix @xmath48 and @xmath66 .
the @xmath67-signature of @xmath54 is the word consisting of the symbols @xmath68 and @xmath69 given by @xmath70 the reduced @xmath67-signature of @xmath54 is the subword of the @xmath67-signature of @xmath54 , given by the repeated removal of adjacent symbols @xmath71 ( in that order ) ; it has the form @xmath72 if @xmath73 then @xmath74 ; otherwise @xmath75 where the rightmost symbol @xmath69 in the reduced @xmath67-signature of @xmath54 comes from @xmath76 .
similarly , if @xmath77 then @xmath78 ; otherwise @xmath79 where the leftmost symbol @xmath68 in the reduced @xmath67-signature of @xmath54 comes from @xmath76 .
it is not hard to verify that this well - defines the structure of a @xmath7-crystal with @xmath80 and @xmath81 in the above notation , with weight function @xmath82 this tensor construction is easily seen to be associative .
the case of two tensor factors is given explicitly by @xmath83 and @xmath84 for a subset @xmath85 define the @xmath86-components of a @xmath7-crystal @xmath45 to be the connected components of the crystal graph using only edges labeled with elements of @xmath86 .
say that @xmath53 is a @xmath86-highest weight vector if @xmath87 for all @xmath88 .
a highest weight vector is an @xmath42-highest weight vector .
let @xmath89 .
for @xmath90 let @xmath91 be the crystal basis of the irreducible integrable @xmath7-module of highest weight @xmath92 .
by @xcite @xmath91 has the structure of a @xmath7-crystal in the sense of section [ subsec : crystals ] where @xmath93 and @xmath94 are the modified chevalley generators .
@xmath91 is connected and has a unique highest weight vector , denoted @xmath95 ; it is the unique element in @xmath91 of weight @xmath92 .
the notion of a dual crystal is given in ( * ? ? ?
* section 7.4 ) .
let @xmath45 be a @xmath7-crystal .
then there is a @xmath7-crystal denoted @xmath96 obtained from @xmath45 by reversing arrows .
that is , @xmath97 with @xmath98 [ pp : dual tensor ] @xcite there is an isomorphism @xmath99 given by @xmath100 .
let @xmath101 be the weyl group of @xmath8 , @xmath102 the simple reflections in @xmath101 .
let @xmath45 be the crystal graph of an integrable @xmath7-module .
say that @xmath103 is an extremal vector of weight @xmath104 provided that @xmath105 and there exists a family of elements @xmath106 such that 1 .
@xmath107 for @xmath108 .
2 . if @xmath109 then @xmath78 and @xmath110 .
3 . if @xmath111 then @xmath74 and @xmath112 .
we now fix notation for the affine case , which , for the most part , follows @xcite .
let @xmath8 be an affine kac - moody algebra over @xmath113 .
let @xmath42 be the index set for the dynkin diagram , @xmath41 the weight lattice , @xmath114 , @xmath115 the natural pairing @xmath116 , @xmath117 the simple roots , @xmath118 the simple coroots , @xmath119 the fundamental weights , @xmath120 the canonical central element , and @xmath121 the null root .
one has @xmath122 .
let the distinguished vertex in the dynkin diagram be labeled @xmath123 as specified in @xcite .
let @xmath124 be the simple lie algebra whose dynkin diagram is @xmath125 and @xmath126 the derived subalgebra .
denote the weight lattices of @xmath127 by @xmath128 , @xmath129 , and @xmath41 respectively .
there are natural projections @xmath130 @xmath131 and @xmath132 .
we shall identify @xmath129 and @xmath128 with the sublattices of @xmath41 given by @xmath133 and @xmath134 where @xmath135 .
let @xmath136 .
denote by @xmath31 , @xmath11 , and @xmath7 the quantized universal enveloping algebras of @xmath32 , @xmath137 , and @xmath8 respectively @xcite .
let @xmath101 and @xmath138 be the weyl groups of @xmath8 and @xmath32 ; they are generated by the simple reflections @xmath102 and @xmath139 respectively .
following @xcite , say that a @xmath11-crystal @xmath45 is _ simple _ if 1 .
@xmath45 is the crystal basis of a finite dimensional integrable @xmath11-module .
2 . there is a weight @xmath140 such that @xmath45 has a unique vector ( denoted @xmath141 ) of weight @xmath92 , and the weight of any extremal vector of @xmath45 is contained in @xmath142 . in the definition of simple crystal in @xcite ,
condition 1 is not present .
however we always want to assume both conditions , so it is convenient to include condition 1 in the definition above .
[ thm : simple ] @xcite 1 .
simple crystals are connected .
the tensor product of simple crystals is simple .
[ rem : simple ] suppose @xmath143 is an isomorphism of simple @xmath11-crystals .
then @xmath144 is the only such isomorphism .
since @xmath144 must preserve weight and send extremal vectors to extremal vectors , it follows that @xmath145 .
since @xmath45 is connected , the rest of the map @xmath144 is determined by the requirement that @xmath144 be a @xmath11-crystal morphism .
suppose @xmath146 are simple @xmath11-crystals . by theorem [ thm : simple ] @xmath147 is simple , with @xmath148 for the @xmath11-crystal @xmath45 , define @xmath149 by @xmath150 define the level of @xmath45 by @xmath151 say that @xmath53 is minimal if @xmath152 .
let @xmath45 be simple .
say that @xmath45 is perfect of level @xmath153 if @xmath154 and @xmath155 and @xmath156 are bijections from the set of minimal elements of @xmath45 to @xmath157 .
the following two results essentially appear in @xcite in the case that @xmath158 .
they are a consequence of the existence of a combinatorial analogue of the @xmath37-matrix on the affinizations of finite @xmath11-crystals and the fact that it satisfies the yang - baxter equation .
[ thm : two tensor ] ( * ? ? ?
* section 4 ) suppose @xmath159 and @xmath160 are simple @xmath11-crystals
. then 1 .
there is a unique isomorphism of @xmath11-crystals @xmath161 .
2 . there is a function @xmath162 , unique up to global additive constant , such that @xmath163 is constant on @xmath86-components and , for all
@xmath164 and @xmath165 , if @xmath166 , then @xmath167 we shall call the maps @xmath168 and @xmath163 the local isomorphism and local energy function on @xmath169 .
the pair @xmath170 is called the combinatorial @xmath37-matrix . by remark [ rem : simple ] and , @xmath171 it is convenient to normalize the local energy function @xmath163 by requiring that @xmath172 with this convention it follows by definition that @xmath173 as operators on @xmath169 .
the following observation , which follows from the uniqueness of the local isomorphism and energy function , is obvious but useful .
[ pp : local r iso ] suppose @xmath159 , @xmath160 , and @xmath174 are simple @xmath11-crystals and there is a @xmath11-crystal isomorphism @xmath175 .
then @xmath176 suppose @xmath146 is a simple @xmath11-crystal for @xmath177 . by abuse of notation
let @xmath178 and @xmath179 denote the local isomorphism and local energy function acting in the @xmath180-th and @xmath181-st tensor positions ( from the right ) .
[ thm : three tensor ] @xcite there is a unique @xmath11-crystal isomorphism @xmath182 given by either side of the yang - baxter equation @xmath183 moreover @xmath184 observe that and are equivalent : the latter is obtained from the former by composing on the right by @xmath185 , using , and the identities @xmath186 and @xmath187 which hold by .
let @xmath146 be a simple @xmath11-module for @xmath177 and let @xmath147 .
let @xmath188 be a sequence of indices in the set @xmath189 .
let @xmath190 denote the permutation that exchanges the @xmath191-th and @xmath192-st positions .
define @xmath193 , the permutation of the set @xmath194 associated with the sequence @xmath195 , by @xmath196 denote by @xmath197 the @xmath11-crystal isomorphism given by @xmath198 and the energy function @xmath199 by @xmath200 [ rem : word iso ] with the above notation , let @xmath201 be a tensor product obtained from @xmath45 by reordering the factors .
then there is a unique @xmath11-crystal isomorphism @xmath202 , and @xmath203 for any sequence @xmath195 such that @xmath204 . by theorem [ thm : simple ] , @xmath45 and @xmath201 are simple since they are tensor products of simple crystals
let @xmath195 be any sequence such that @xmath204 .
then @xmath205 is an isomorphism of simple @xmath11-crystals .
but there is a unique @xmath11-crystal isomorphism @xmath206 by remark [ rem : simple ] .
[ rem : exchange energy ] if @xmath207 is obtained from @xmath195 by a sequence of exchanges of the form @xmath208 where @xmath209 , then it is easy to see that @xmath210 . in general @xmath211 .
take @xmath212 in @xmath213 with @xmath214 ( see sections [ sec : crystal a ] and [ sec : a perf ] ) . then @xmath215 and @xmath216 . for later use
we define a few specific reduced words .
let @xmath217 note that @xmath218 is the permutation that reverses the numbers @xmath194 .
for @xmath153 and @xmath219 positive integers summing to @xmath220 , define @xmath221 observe that @xmath222 is the shortest permutation that exchanges the numbers @xmath223 past the numbers @xmath224 .
let @xmath153 and @xmath219 be positive integers .
let @xmath146 be a simple @xmath11-crystal for @xmath225 .
then @xmath226 and @xmath227 are simple crystals by theorem [ thm : simple ] .
we compute the local isomorphism @xmath228 and energy function @xmath229 explicitly .
[ pp : weave ] with the above notation , @xmath230 proposition [ pp : weave ] is proved by induction .
the first nontrivial cases are for three tensor factors .
[ pp : rh 3 ] @xmath231 we will prove the case @xmath232 and @xmath233 as the case @xmath234 and @xmath235 is similar .
we have @xmath236 .
it must be shown that @xmath237 .
let us fix @xmath238 .
for any isomorphism @xmath144 that reorders @xmath239 by a composition of local isomorphisms , define @xmath240 to be the position in the threefold tensor @xmath241 where @xmath242 acts ( see section [ sec : tensor crystal ] ) .
let @xmath243 , @xmath244 , @xmath245 .
write @xmath246 and let @xmath247 ( resp .
@xmath248 ) have value @xmath220 ( left ) and @xmath37 ( right ) on the two - fold tensor product @xmath249 ( resp .
@xmath250 ) .
then the rows of the following table give all the possibilities : @xmath251 the columns @xmath247 , @xmath248 , and @xmath252 agree with the defining conditions for @xmath163 in theorem [ thm : two tensor ] .
the next case is @xmath233 .
[ pp : rh ell 1 ] @xmath253 formula holds by remark [ rem : simple ] .
formula is proved by induction on @xmath153 , by applying and for @xmath254 where @xmath255 , @xmath256 , and @xmath257
. again follows by remark [ rem : simple ] .
the proof of follows by induction on @xmath219 .
the base of the induction ( @xmath233 ) is given by proposition [ pp : rh ell 1 ] .
for the induction step , and are applied for @xmath258 where @xmath259 , @xmath260 and @xmath261 .
let @xmath45 be a simple @xmath11-crystal , equipped with a function @xmath262 called its intrinsic energy , which is required to be constant on @xmath86-components and defined up to a global additive constant .
call the pair @xmath263 a graded simple @xmath11-crystal .
we normalize the intrinsic energy function by the requirement that @xmath264 let @xmath265 be graded simple @xmath11-crystals for @xmath266 .
write @xmath267 for the projection onto the rightmost tensor factor .
define @xmath268 by @xmath269 from now on we suppress the operator @xmath267 , observing that the operator @xmath270 always acts on the rightmost tensor factor , which will be in @xmath146 .
the subscripts of @xmath163 and @xmath168 indicate positions .
observe that holds for @xmath271 if it is assumed that it holds for @xmath159 and @xmath160 and that holds .
this can be seen using and .
[ pp : graded tensor ] graded simple crystals form a tensor category .
it must be shown that the tensor product construction for graded simple crystals is associative .
this is true if the grading is ignored .
it suffices to show that @xmath272 by proposition [ pp : rh 3 ] and we have @xmath273 and @xmath274 in the second computation , is used .
also used is @xmath275 , which holds because @xmath270 acts on the rightmost tensor factor , which is not changed by @xmath276 .
one sees that and are equal by .
let @xmath265 be graded simple @xmath11-crystals for @xmath277 and let @xmath278 .
let @xmath147 .
following @xcite define the energy function @xmath279 by @xmath280 by the normalization assumption it follows that @xmath281 the following formula was motivated by the definition of the @xmath282 energy function in ( * ? ? ?
* section 3.3 ) .
[ pp : intrinsic gen ] the intrinsic energy @xmath283 for the @xmath220-fold tensor product @xmath147 is given by @xmath284 the proof proceeds by induction on @xmath220 . for @xmath285
holds by definition .
suppose it holds for @xmath286 tensor factors .
let @xmath287 and @xmath288 . applying the definition for @xmath289 , induction , and proposition [ pp : weave ] , we have @xmath290 which is evidently the right hand side of .
the energy functions are insensitive to reordering of tensor factors .
[ pp : reorder de ] let @xmath201 be a reordering of @xmath291 and @xmath202 any composition of local isomorphisms
. then @xmath292 and @xmath293 .
the proof immediately reduces to the case @xmath294 .
let us prove the first assertion , as the second follows from it .
write @xmath295 , @xmath296 , @xmath297 and @xmath298 . by
it follows that @xmath299 where @xmath300 . therefore @xmath301 as graded simple crystals .
tensoring both on the left by @xmath302 and on the right by @xmath303 , the result is a pair of isomorphic graded simple crystals @xmath304 via the map @xmath305 . in particular
this argument implicitly uses proposition [ pp : graded tensor ] and remark [ rem : simple ] .
we recall conjectures regarding the crystals @xmath17 .
recall chevalley s partial order on @xmath128 , defined by @xmath307 if and only if @xmath308 .
[ conj : brs ] @xcite for each @xmath309 and @xmath310 , there exists an irreducible finite dimensional integrable @xmath11-module @xmath311 with simple crystal basis @xmath17 having a unique extremal vector @xmath312 of weight @xmath313 , and a prescribed @xmath31-crystal decomposition of the form @xmath314 , where @xmath45 is a direct sum of @xmath31-crystals of the form @xmath91 where @xmath140 and @xmath315 .
moreover there is a prescribed intrinsic energy function @xmath316 , that is constant on @xmath86-components , such that @xmath317 and @xmath54 is any element not in the @xmath86-component of @xmath312 .
@xcite @xmath311 is the universal @xmath11-module generated by an extremal vector of weight @xmath318 .
this is known to hold for @xmath18 @xcite .
[ conj : grade ] @xcite there is a unique @xmath319 such that @xmath320 .
moreover @xmath321 where @xmath322 is the local energy function .
[ rem : grade ] observe that the function @xmath323 given by @xmath324 is constant on @xmath86-components . by definition @xmath325 for @xmath326 . by ,
@xmath327 for all @xmath328 .
the assertion follows from the fact that @xmath163 is constant on @xmath86-components of @xmath329 , by theorem [ thm : two tensor ] .
let @xmath330 and @xmath331 where @xmath332 for @xmath333 .
the set @xmath334 of ( classically restricted ) paths in @xmath45 of weight @xmath92 is the set of @xmath86-highest weight vectors in @xmath45 of weight @xmath92 .
define the classically restricted one dimensional sum @xmath335
in this section we discuss the combinatorial properties of type @xmath34 in detail . in this section
let @xmath32 be the complex simple lie algebra of type @xmath336 .
let @xmath337 be the index set for the vertices of its dynkin diagram .
let us realize the weight lattice @xmath128 explicitly inside the real span of the hyperplane in @xmath338 orthogonal to the vector @xmath339 .
let @xmath340 be the standard basis of @xmath338 .
then one may take @xmath341 for @xmath326 . identifying the weight lattice of @xmath342 with @xmath343
, there is a projection @xmath344 defined by @xmath345 where @xmath346 is the standard bilinear form on @xmath338 . in coordinates it is given by @xmath347 .
in particular @xmath348 for @xmath326 .
the highest root of @xmath32 is @xmath349 . if @xmath92 is a partition @xmath350 then @xmath351 . by abuse of notation an element of @xmath343
is sometimes identified with its image under @xmath352 .
recall that for @xmath140 , @xmath91 is the crystal of the irreducible integrable @xmath31-module of highest weight @xmath92 .
we now review the explicit realization of @xmath91 given in @xcite .
the vertices of @xmath91 are given by the set of ( semistandard ) tableaux of shape @xmath92 in the alphabet @xmath353=\{1,2,\dots , n\}$ ] .
the @xmath31-crystal structure shall be defined by embedding tableaux into the set of words .
the crystal graph of @xmath354 ( where recall that @xmath355 is the image under @xmath352 of the partition @xmath356 ) is given by the set @xmath353 $ ] with @xmath357 for all @xmath326 and all other elements sent to @xmath358 .
one defines @xmath359 for all @xmath360 $ ] .
the set of words in the alphabet @xmath353 $ ] is the crystal graph of the tensor algebra of @xmath354 ; its explicit structure is given by the signature rule ( see section [ sec : tensor crystal ] ) .
the weight function @xmath361 on this crystal graph is given as follows .
let the content of a word @xmath362 be the element @xmath363 where @xmath364 is the number of occurrences of the letter @xmath191 in @xmath362
. then @xmath365 .
we identify the tableau @xmath366 with its column - reading word , which is by definition @xmath367 where @xmath368 is the strictly decreasing word given by the @xmath191-th column of @xmath369 .
this defines an embedding of @xmath91 into the set of words in the alphabet @xmath353 $ ] .
it is easy to see that @xmath93 and @xmath94 stabilize the image of this embedding .
the @xmath31-crystal structure on @xmath91 is defined by declaring this embedding to be a morphism of @xmath31-crystals .
the highest weight vector @xmath95 of @xmath91 is given explicitly by the yamanouchi tableau , the tableau of shape @xmath92 whose @xmath67-th row is filled with the letter @xmath67 for all @xmath67 .
schensted @xcite gave an insertion algorithm that associates to a word @xmath362 a tableau @xmath370 of partition shape .
given an alphabet ( totally ordered set ) @xmath371 , let @xmath372 be the equivalence relation on words in the alphabet @xmath371 defined by @xmath373 if and only if @xmath374 .
let @xmath375 denote the set of @xmath372-classes .
it follows from @xcite that @xmath376 , so that the multiplication on words given by juxtaposition , descends to a multiplication in @xmath375 .
@xmath375 is called the plactic monoid .
knuth @xcite showed that the set of relations in @xmath375 is generated by relations of the following form , where @xmath377 and @xmath378 are words : @xmath379 a column word is one that is strictly decreasing .
let @xmath380 be a subinterval .
for a word @xmath362 in the alphabet @xmath371 , denote by @xmath381 the subword of @xmath362 obtained by erasing all of the letters not in @xmath45 .
it follows immediately from the relations that if @xmath373 then @xmath382 . it is a well - known and remarkable fact that schensted s algorithm gives a morphism of type @xmath336 crystal graphs .
[ thm : crystal p ] for all words @xmath362 in the alphabet @xmath353 $ ] and all @xmath326 , @xmath383 .
it suffices to show that if @xmath373 is one of the relations then @xmath384 is also .
the result is entirely straightforward unless @xmath385 and @xmath93 changes one of @xmath386 . in this case
( ignoring identical left and right factors ) one has @xmath387 .
applying @xmath93 to both sides one obtains @xmath388 .
let @xmath8 be the affine algebra of type @xmath389 .
write @xmath390 .
[ thm : a perf ] @xcite for every @xmath309 and @xmath24 there is a finite dimensional integrable @xmath391-module @xmath311 with crystal basis @xmath17 such that 1 . @xmath392 as @xmath393-crystals .
2 . @xmath394 .
@xmath17 is perfect of level @xmath395 .
moreover , let @xmath396 be the rotation of the dynkin diagram of type @xmath389 that sends the simple root @xmath397 to @xmath398 where subscripts are taken modulo @xmath399 .
then there is a unique bijection @xmath400 such that @xmath401 where @xmath402 is induced by the map @xmath403 given by @xmath404 .
the operators @xmath242 and @xmath405 are computed explicitly in @xcite .
it suffices to define @xmath406 . for @xmath407
define @xmath408 by the property @xmath409}={\mathbb{p}}(t|_{[2,n]})-1\ ] ] where @xmath410 is the tableau obtained by adding the integer @xmath411 to each letter .
this definition determines the positions in @xmath412 of the letters in the subinterval @xmath413 $ ] .
but it also determines the positions of the letters @xmath399 as they must fill up the rest of the rectangular shape .
let @xmath414 , @xmath415 , @xmath416 .
then a tableau @xmath407 and @xmath412 are given below .
} = \begin{array}{cccc } \hline { \multicolumn{1}{|c| } { } } & { \multicolumn{1}{c| } { } } & { \multicolumn{1}{c|}{2 } } & { \multicolumn{1}{c|}{3 } } \\ \cline{1 - 4 } { \multicolumn{1}{|c|}{2 } } & { \multicolumn{1}{c|}{2 } } & { \multicolumn{1}{c|}{3 } } & { \multicolumn{1}{c|}{4 } } \\ \cline{1 - 4 } { \multicolumn{1}{|c|}{3 } } & { \multicolumn{1}{c|}{4 } } & { \multicolumn{1}{c|}{5 } } & { \multicolumn{1}{c|}{5 } } \\ \cline{1 - 4 } \end{array } \qquad { \mathbb{p}}(t|_{[2,n ] } ) = \begin{array}{cccc } \hline { \multicolumn{1}{|c|}{2 } } & { \multicolumn{1}{c|}{2 } } & { \multicolumn{1}{c|}{2 } } & { \multicolumn{1}{c|}{3 } } \\ \cline{1 - 4 } { \multicolumn{1}{|c|}{3 } } & { \multicolumn{1}{c|}{3 } } & { \multicolumn{1}{c|}{4 } } & { \multicolumn{1}{c|}{5 } } \\ \cline{1 - 4 } { \multicolumn{1}{|c|}{4 } } & { \multicolumn{1}{c|}{5 } } &
{ \multicolumn{1}{c|}{}}&{\multicolumn{1}{c|}{}}\\ \cline{1 - 4 } \end{array } \\ \\
{ \mathbb{p}}(t|_{[2,n]})-1 = \begin{array}{cccc } \hline { \multicolumn{1}{|c|}{1 } } & { \multicolumn{1}{c|}{1 } } & { \multicolumn{1}{c|}{1 } } & { \multicolumn{1}{c|}{2 } } \\ \cline{1 - 4 } { \multicolumn{1}{|c|}{2 } } & { \multicolumn{1}{c|}{2 } } & { \multicolumn{1}{c|}{3 } } & { \multicolumn{1}{c|}{4 } } \\ \cline{1 - 4 } { \multicolumn{1}{|c|}{3 } } & { \multicolumn{1}{c|}{4 } } & { \multicolumn{1}{c|}{}}&{\multicolumn{1}{c|}{}}\\ \cline{1 - 4 } \end{array } \qquad { { \psi}}^{-1}(t)= \begin{array}{cccc } \hline { \multicolumn{1}{|c|}{1 } } & { \multicolumn{1}{c|}{1 } } & { \multicolumn{1}{c|}{1 } } & { \multicolumn{1}{c|}{2 } } \\ \cline{1 - 4 } { \multicolumn{1}{|c|}{2 } } & { \multicolumn{1}{c|}{2 } } & { \multicolumn{1}{c|}{3 } } & { \multicolumn{1}{c|}{4 } } \\ \cline{1 - 4 } { \multicolumn{1}{|c|}{3 } } & { \multicolumn{1}{c|}{4 } } & { \multicolumn{1}{c|}{5 } } & { \multicolumn{1}{c|}{5 } } \\ \cline{1 - 4 } \end{array } \end{split}\ ] ] since @xmath17 is perfect of level @xmath395 , there is a unique @xmath418 such that @xmath419 ; it is the tableau in @xmath17 whose @xmath67-th row from the bottom consists of @xmath395 copies of the value @xmath420 . since @xmath17 consists of a single @xmath86-component , conjecture [ conj : grade ] holds by remark [ rem : grade ] .
let @xmath421 be a sequence of rectangular partitions .
say that @xmath422 has @xmath423 rows and @xmath190 columns where @xmath424 and @xmath425 for all @xmath426 .
define the @xmath11-crystal @xmath427 the elements of @xmath428 are called inhomogeneous paths . of course one could compute @xmath242 and @xmath405 on @xmath428 using the signature rule and the above rule for @xmath242 and @xmath405 on single tensor factors , but it is simpler to compute them as follows
let @xmath429 be the bijection given by @xmath430 .
then @xmath242 and @xmath405 as operators on @xmath428 may be defined by the equations and .
consider the finite dimensional integrable @xmath11-module @xmath311 with crystal basis @xmath17 ( see theorem [ thm : a perf ] ) .
we observe that the crystal @xmath17 is simple ( see section [ sec : simple crystal ] ) .
let @xmath431 , with corresponding partition @xmath432 .
it is not hard to check that the extremal vectors of @xmath17 are in bijection with the orbit @xmath142 . in the notation of section [ sec : extremal vector ] , for every @xmath433 ,
let @xmath434 be the unique tableau in @xmath17 of content @xmath435 .
this shows that @xmath17 is simple .
the crystal @xmath428 of the previous section is a tensor product of simple crystals .
so @xmath428 is simple and hence connected .
let @xmath436 and @xmath437 . for a column word @xmath120 in the alphabet
@xmath353 $ ] , let @xmath438 be the column word that uses precisely those letters that are not in @xmath120 .
for a sequence of column words @xmath439 , define @xmath440 [ pp : dual irr ] there is a unique @xmath31-crystal isomorphism @xmath441 .
it is easy to see that the map @xmath442 given by @xmath443 ( where @xmath444 is defined in ) is a well - defined bijection that satisfies . for uniqueness , by definition
the map @xmath445 takes highest weight vectors to lowest weight vectors . but
@xmath91 has a unique highest weight vector and @xmath446 has a unique lowest weight vector .
[ pp : dual p ] let @xmath447 where each @xmath368 is a column word
. then @xmath448 . in the statement , @xmath449
is computed with respect to a factorization into @xmath450 column words , the last several of which may be empty .
it suffices to show that if @xmath451 are column words such that @xmath452 with @xmath453 one letter longer than @xmath362 , then @xmath454 .
this reduction , together with the proof of this special case , can both be seen by considering a jeu de taquin on skew tableaux having at most two columns .
let @xmath455 .
one has @xmath456 and dually @xmath457 .
[ pp : dual perfect ] the @xmath11-crystal @xmath17 has dual crystal @xmath458 . observe that as @xmath31-crystals one has @xmath459 using @xmath460 columns . by the uniqueness in proposition [ pp : dual irr ]
it follows that the crystals are also dual as @xmath11-crystals by the same map .
[ thm : flip ] let @xmath45 be the crystal basis of a finite dimensional @xmath31-module ( resp .
@xmath11-module ) . then there is an involution @xmath461 denoted @xmath462 , such that @xmath463 for all @xmath326 ( resp .
@xmath48 with subscripts taken modulo @xmath399 ) where @xmath464 is the longest element of @xmath138 .
consider the orientation - reversing automorphism of the dynkin diagram of type @xmath389 that sends @xmath397 to @xmath465 for all @xmath67 with subscripts taken modulo @xmath399 .
the existence of this automorphism implies the existence of a map @xmath466 satisfying .
since the dynkin diagram automorphism is an involution , so is the induced map @xmath467 .
the following result is easily verified .
[ pp : flip tensor ]
suppose there are crystal graphs @xmath146 and bijections @xmath468 satisfying .
let @xmath331 and @xmath469 .
then the map @xmath470 given by @xmath471 satisfies .
given a word @xmath362 , let @xmath472 be the word obtained by replacing each letter @xmath67 by @xmath420 , and reversing the resulting word .
clearly if @xmath362 is a column word then so is @xmath472 . if @xmath473 where @xmath368 is a column word for all @xmath191 , then by definition @xmath474 , which is a sequence of column words . in particular , let @xmath366 .
it is easy to check that @xmath475 is a skew tableau whose shape is given by the 180 degree rotation of @xmath92 .
define @xmath476 .
the tableau @xmath477 is called the evacuation @xcite of the tableau @xmath369 .
observe that if @xmath92 is a rectangle then @xmath478 .
[ pp : flip p ] 1 .
the map @xmath479 is an involution satisfying . 2 .
for any word @xmath362 , @xmath480 .
the map @xmath481 given by @xmath482 is the unique map satisfying ( in which @xmath483 is replaced by @xmath484 ) .
part 1 is straightforward . for part 2
it suffices to show that if @xmath373 is a relation of the form then so is @xmath485 , but this is also straightforward .
part 3 follows from part 2 , the connectedness of @xmath91 , and the fact that @xmath486 is a morphism of crystal graphs .
we remark that the uniqueness in part 3 , together with theorem [ thm : flip ] , imply that @xmath482 is an involution .
[ pp : dual flip ] there is a bijection @xmath487 given by @xmath488 such that @xmath489 for all @xmath366 and @xmath326 .
moreover if @xmath92 is a rectangle then also holds for @xmath490 .
@xmath491 and @xmath467 obviously commute on sequences of column words .
proposition [ pp : dual p ] implies that @xmath492 for all @xmath366 .
equation follows from and theorem [ thm : flip ] .
a similar statement holds for tensor products .
let @xmath332 for @xmath266 . the existence and uniqueness of the local isomorphism @xmath493 ( see section [ sec : comb r def ] ) is guaranteed by the simplicity of @xmath17 ( see section [ sec : a simple ] ) .
observe that as a @xmath31-crystal @xmath169 is multiplicity - free , since @xmath494 is indexed by a rectangular partition for @xmath266 @xcite .
therefore the isomorphism @xmath495 is uniquely specified by the property that @xmath496 is the unique element @xmath497 such that @xmath498 .
the bijection @xmath168 is described quite explicitly in @xcite .
let @xmath499 then @xmath500 hence @xmath501 [ pp : r dual ] the following diagram commutes : @xmath502 let @xmath65 for @xmath266 and @xmath503 .
then @xmath504 by the definition of @xmath168 and proposition [ pp : dual p ] .
this suffices by the definition of @xmath168 .
[ pp : r ev ] the following diagram commutes : @xmath505 let @xmath506 be as before for @xmath266 .
@xmath507 by the definition of @xmath168 and proposition [ pp : flip p ] .
this suffices by the definition of @xmath168 . in the case that @xmath508 for @xmath266
there is an explicit construction for the local isomorphism @xmath168 which will be useful later . for @xmath509 , represent @xmath510 by a column of height @xmath399 with a dot at height @xmath67 if and only if the letter @xmath67 occurs in @xmath54 .
now @xmath511 with @xmath512 is obtained as follows @xcite .
let @xmath513 and @xmath514 . 1 .
pick the highest dot in @xmath515 . call it @xmath516 .
connect @xmath516 with the highest dot in @xmath517 not higher than @xmath516 ( assuming periodic boundary conditions if necessary ) .
2 . repeat this with all unconnected dots in @xmath515 .
3 . slide all @xmath518 unpaired dots from @xmath517 to @xmath515 .
the result is @xmath519 .
let @xmath520 , @xmath521 and @xmath522 .
then @xmath523 so that @xmath524 and @xmath525 .
explicit formulas may be given for the energy functions in type @xmath371 .
let @xmath332 with @xmath86-highest weight vector @xmath526 for @xmath266 .
given the normalization one has the explicit formula @xcite @xmath527 where @xmath528 is the number of cells of @xmath92 that lie in columns strictly to the right of the @xmath395-th column .
it is not hard to see that the values of @xmath163 are nonpositive integers .
[ rem : different energy ] in @xcite a different normalization was used , namely , only the @xmath529 function was present .
let @xmath530 and @xmath531 where @xmath428 is defined in .
write @xmath532 and @xmath533 for @xmath534 and @xmath535 in and . since @xmath394 , @xmath536 [ pp : energy dual ] for all @xmath537 , @xmath538 let @xmath539 , @xmath201 be any reordering of the tensor factors of @xmath428 , and @xmath202 any composition of local isomorphisms
. by propositions [ pp : r dual ] and [ pp : r ev ] , @xmath540 using this one may reduce to the case @xmath541 .
the crystal @xmath428 is connected ( see section [ sec : a simple ] ) .
it then suffices to check for the single element @xmath542 where @xmath526 is the @xmath86-highest weight vector of @xmath146 , by and proposition [ pp : dual flip ] .
again by proposition [ pp : dual flip ] @xmath543 is a @xmath86-highest weight vector of the same weight ( namely @xmath544 ) as @xmath545 where @xmath546 is the unique @xmath86-highest weight vector in @xmath547 .
but there is only one such @xmath86-highest weight vector in @xmath548 of this weight , so the two vectors must agree .
it follows that both sides of evaluate to zero .
let @xmath92 be a partition ( or an element of @xmath549 via @xmath352 as in section [ sec : root a ] ) , @xmath550 a sequence of rectangles and @xmath428 as in .
define @xmath551 with @xmath552 as in and @xmath553 it was shown in @xcite ( see also @xcite ) that @xmath554 is the kostka polynomial @xmath555 if @xmath37 is a sequence of single rowed partitions where @xmath556 and @xmath557 .
hence we call @xmath554 the generalized _ kostka polynomial_. both @xmath554 and @xmath558 have nonnegative integer coefficients .
the @xmath86-components of @xmath428 are parametrized by the combinatorial objects called littlewood
richardson ( lr ) tableaux . these objects , being in the multiplicity space , should be considered a completely different kind of object than the elements of @xmath428 .
it is a beautiful coincidence in type @xmath371 that operators that act on the representation space such as crystal operators and duality , also act in the multiplicity space in certain situations .
we recall the notion of an @xmath37-lr word @xcite , which can be formulated as follows .
let @xmath559 be a sequence of positive integers summing to @xmath450 , @xmath560 , @xmath561 the subinterval of @xmath371 given by the first @xmath562 numbers , @xmath563 the next @xmath564 numbers , etc .
say that a word @xmath362 in the alphabet @xmath371 is @xmath565-balanced if @xmath566 for all @xmath567 that are not maximum in one of the subintervals @xmath568 .
let @xmath530 be a sequence of rectangular partitions such that @xmath422 has @xmath569 rows and @xmath570 columns .
let @xmath571 be obtained by juxtaposing the parts of the @xmath422 .
define the set @xmath572 of @xmath37-lr words by @xmath573 if and only if @xmath362 has content @xmath574 and @xmath362 is @xmath565-balanced .
this definition is equivalent to the one in @xcite .
conversely , for an @xmath565-balanced word @xmath362 , let @xmath575 be the content of @xmath362 .
then @xmath576 for all @xmath577 in the same subalphabet @xmath568 .
let @xmath570 be this common value for the subalphabet @xmath568 and @xmath422 be the rectangular partition having @xmath569 rows and @xmath570 columns .
then @xmath362 is an @xmath37-lr word .
let @xmath578 be the subset of @xmath572 consisting of tableaux of partition shape ( via the identification of a tableau with its column - reading word ) and @xmath579 the subset of @xmath578 consisting of tableaux of partition shape @xmath92 .
the set @xmath579 is empty unless @xmath92 has at most @xmath450 parts .
the way in which the lr tableaux parametrize the multiplicity space of @xmath428 , is by suitable recording tableaux for schensted s insertion .
consider a given factorization @xmath580 of @xmath362 where @xmath368 is a column word in the alphabet @xmath353 $ ] .
let @xmath581 be the unique filling of the ferrers diagram of the shape of @xmath582 such that the shape of @xmath583}$ ] is the shape of @xmath584 for all @xmath191 .
since each @xmath368 is a column word it follows that the transpose @xmath585 of @xmath586 is a semistandard tableau . in more traditional language
, @xmath586 is the recording tableau for the column insertion of the word @xmath587 such that the insertion of the letters in @xmath368 are recorded by the letter @xmath191 .
fix @xmath588 and let @xmath589 where @xmath590 denotes transpose .
let @xmath399 be a positive integer such that @xmath591 for all @xmath191 .
one may regard elements of @xmath592 as tableaux of a fixed skew shape in the alphabet @xmath353 $ ] .
the following result , up to labeling of the recording tableaux , is a special case of the main theorem in @xcite .
[ path lr ] there is a bijection @xmath593 given by @xmath594 where @xmath92 runs over partitions with @xmath595 .
recall that the operation @xmath491 was defined for tableaux @xmath369 in the alphabet @xmath353 $ ] with at most @xmath450 columns and for partitions with at most @xmath399 rows and @xmath450 columns .
we define an operation @xmath596 which is a similar kind of operation but with @xmath399 and @xmath450 exchanged .
let @xmath597 be the partition obtained by skewing the rectangle @xmath598 by @xmath92 and rotating 180 degrees , or equivalently @xmath599 .
let @xmath600 where @xmath601 for each @xmath191 .
this complements the widths of the rectangles in @xmath37 . for a sequence @xmath362 of @xmath399 column words in the alphabet
@xmath602 $ ] , we write @xmath603 for the dual sequence of column words instead of @xmath604 .
[ pp : record dual ] @xmath605 let @xmath606 and @xmath607 be the tableaux on the left and right hand sides .
let @xmath608 be obtained from @xmath362 by replacing @xmath609 by @xmath358 and let @xmath610 and @xmath611 be the corresponding tableaux for @xmath608 . by induction on the number of nonempty factors @xmath612 .
by definition the first @xmath613 column words of @xmath614 and @xmath615 agree . to conclude @xmath616 it suffices to show : 1 .
@xmath617}$ ] for @xmath618 .
2 . @xmath606 and @xmath607 have the same shape .
part 1 holds by the definition of @xmath619 . for part 2
one has @xmath620 by the fact that the @xmath486 and @xmath619 tableaux of a given word have the same shape , propositions [ pp : flip p ] and [ pp : dual p ] , the fact that @xmath621 is shape - preserving , and the rule for how @xmath491 and @xmath596 change shapes . there is a bijection @xmath622 given by @xmath623 .
let @xmath624 .
then @xmath625 is a tableau of shape @xmath597 by proposition [ pp : dual irr ] .
it has the correct content to be @xmath626-lr by definition since the corresponding rectangles in @xmath37 and @xmath626 have the same heights .
since @xmath369 is @xmath565-balanced , @xmath625 is also , by .
thus the desired map is well - defined .
since @xmath596 is an involution the map is bijective .
let @xmath627 and @xmath628 and @xmath416 .
then @xmath629 and @xmath37 and @xmath626 ( whose rectangles have their rows are filled with the letters of their corresponding subintervals ) are given by @xmath630 let @xmath631 . then @xmath632 .
below is a tableau @xmath633 and its dual @xmath634 .
@xmath635 given @xmath636 , let @xmath637 be obtained from @xmath37 by exchanging the @xmath638-th and @xmath639-st rectangles .
there is a bijection @xcite given by the generalized automorphisms of conjugation @xmath640 .
it extends uniquely to a bijection @xmath641 by @xmath642 and @xmath643 .
the bijections @xmath644 and the combinatorial @xmath37-matrices @xmath645 are related as follows ( which may also be used as the definition of @xmath644 ) . [ r auto ] for any @xmath646 and @xmath636 , @xmath647 .
[ gen auto dual ] the following diagram commutes : @xmath648 let @xmath649 . by proposition [ path lr ]
there is a @xmath646 such that @xmath650 .
then @xmath651 by propositions [ pp : record dual ] , [ r auto ] , [ pp : r dual ] and [ pp : r ev ] .
there are embeddings @xmath652 where @xmath653 is the sequence of single - rowed shapes of sizes given by @xmath574 ; they are defined as compositions of two kinds of elementary embeddings @xcite .
the second kind of elementary embedding is given by the @xmath644 .
the first kind is given as follows .
suppose that @xmath654 and @xmath655 where @xmath656 .
write @xmath657 .
then there is an embedding @xmath658 .
this extends to an embedding @xmath659 by @xmath660 where @xmath661 and @xmath662 .
again fix @xmath399 .
we have @xmath663 and @xmath664 .
[ elem embed dual ] the following diagram commutes : @xmath665 one reduces to the two rectangle case by restriction to the subinterval @xmath666 .
recalling that the tensor product of two rectangles is multiplicity - free and using the duality symmetry of tensor product multiplicities , one has @xmath667 the following diagram commutes : @xmath668 this holds by propositions [ gen auto dual ] and [ elem embed dual ] .
let @xmath530 such that @xmath422 has @xmath569 rows and @xmath570 columns .
the generalized charge is a function @xmath669 defined as follows . for @xmath670 @xmath671
is the zero function . for @xmath672 and @xmath673
, let @xmath674 be the number of cells in the shape of @xmath370 to the right of the @xmath675-th column ( see ) . for general @xmath220 and @xmath624 , define @xcite @xmath676 where @xmath677 refers to the above function @xmath529 applied to the shape of the @xmath486 tableau of the restriction of @xmath678 to the @xmath191-th and @xmath192-st subalphabets for @xmath679 .
[ charge energy ] @xcite for all @xmath646 , @xmath680 .
[ charge dual ] for all @xmath681 , @xmath682 .
this follows from propositions [ path lr ] , [ charge energy ] , and [ pp : energy dual ] .
rigged configurations are combinatorial objects which were first introduced in the context of bethe ansatz studies of spin models in statistical mechanics @xcite . here
we recall their definition and a bijection from littlewood
richardson tablaux ( or equivalently classically restricted paths of type @xmath34 ) to rigged configurations .
the relation between duality and this bijection is given in theorem [ cdual rc ] .
we will need these results later in section [ sec : fermi ] in the proof of fermionic formulas of type @xmath21 , @xmath1 and @xmath2 . in section [ sec : fermi ] we will also introduce rigged configurations associated to @xmath21 , @xmath1 and @xmath2 .
the rigged configurations of this section correspond to the algebra @xmath34 .
let @xmath92 be a partition and let @xmath530 be a sequence of rectangles with @xmath422 having @xmath423 rows and @xmath190 columns .
the set of admissible configurations @xmath683 is the set of all sequences of partitions @xmath684 subject to the constraints @xmath685 for @xmath686 and @xmath687 . here
@xmath688 is the vacancy number of the parts ( strings ) of length @xmath67 in @xmath689 defined as @xmath690 where @xmath691 is the empty partition , @xmath692 is the size of the first @xmath67 columns of the partition @xmath693 , and @xmath694 is the partition whose parts are the widths of the rectangles in @xmath37 of height @xmath180 .
the set @xmath695 of rigged configurations is defined as follows .
an element of @xmath696 consists of a configuration @xmath697 together with a rigging @xmath86 .
the rigging @xmath86 is a double sequence of partitions @xmath698 such that @xmath699 is a partition in a box of width @xmath688 and height @xmath700 where @xmath700 is the number of parts of @xmath689 of size @xmath67 .
the cocharge of a rigged configuration @xmath696 is defined by @xmath701 where @xmath702 is the size of the @xmath67-th column of @xmath689 .
a rigged configuration @xmath703 can also be viewed as follows .
each part of @xmath699 labels a part of length @xmath67 in @xmath689 .
the pair @xmath704 of a part of @xmath689 together with its label is called a string .
then @xmath703 is the multiset of strings .
a string @xmath704 is called singular if @xmath705 .
let @xmath92 be a partition and @xmath37 a sequence of rectangles @xmath706 with @xmath707 rows and @xmath708 columns .
fix @xmath399 such that @xmath709 and @xmath710 for all @xmath711 , and let @xmath712 .
[ rc dual ] 1 .
there is a bijection @xmath713 defined by @xmath714 for all @xmath715 .
@xmath716 for all @xmath717 and @xmath715 .
3 . there is an induced bijection @xmath718 defined by @xmath714 and @xmath719 for all @xmath715 and @xmath717 .
4 . @xmath720 .
note that @xmath721 .
since @xmath722 , @xmath723 and @xmath724 it follows that @xmath725 so that indeed @xmath726 .
the other assertions follow easily from the definitions . in ref .
@xcite the existence of several bijections between lr tableaux and rigged configurations was established .
let @xmath727 and @xmath728 be the columnwise quantum and coquantum number bijections , respectively .
let us describe the algorithm for the bijection @xmath729 explicitly .
let @xmath624 be an lr tableau for a partition @xmath92 and a sequence of rectangles @xmath550 where @xmath706 has @xmath707 rows and @xmath708 columns
. set @xmath730 $ ] .
relabel the letters in @xmath369 such that the @xmath67-th occurrence ( from the left ) of the @xmath195-th largest letter in @xmath568 is mapped to @xmath731 .
call the relabelled tableau @xmath732 .
set @xmath733 .
let @xmath734 be the tableau of shape @xmath422 such that the columns of @xmath734 are successively labelled by the letters in @xmath735 . the rigged configuration @xmath736 is obtained recursively by constructing a rigged configuration @xmath737 for each letter @xmath738 occurring in @xmath732 .
set @xmath739 .
suppose that @xmath740 , and denote the column index of @xmath741 in @xmath732 by @xmath120 and the column index of @xmath741 in @xmath734 by @xmath742 .
define the numbers @xmath743 for @xmath744 as follows .
let @xmath745 be the length of the longest singular string in @xmath746 .
now select inductively a singular string in @xmath747 for @xmath748 whose length @xmath743 is maximal such that @xmath749 ; if no such string exists set @xmath750
. then @xmath737 is obtained from @xmath751 by adding one box to the selected strings with labels such that they remain singular , leaving all other strings unchanged .
then the image of @xmath369 under @xmath752 is given by @xmath753 .
for the above algorithm it is necessary to be able to compute the vacancy numbers of an intermediate configuration @xmath754
. suppose @xmath741 occurs in @xmath734 in column @xmath742 .
in general @xmath755}))$ ] is not a sequence of rectangles .
if @xmath756})$ ] is not a rectangle one splits it into two rectangles , one of width @xmath742 and one of width @xmath757 .
the vacancy numbers are calculated with respect to this new sequence of rectangles . by definition
set @xmath758 for @xmath759 and @xmath750 if @xmath760 .
[ ex : bij ] let @xmath761 , @xmath762 .
consider @xmath633 given by @xmath763 the non - trivial steps of the above algorithm for @xmath752 are given in table [ table : ex ] .
a rigged partition is represented by its ferrers diagram where to the right of each part the corresponding rigging is indicated .
the vacancy numbers are given to the left of each part .
for example @xmath764 so that the vacancy numbers of @xmath765 are calculated with respect to the sequence of rectangles @xmath766 . @xmath767 x & ( \nu , j)_{(x)}^{(1 ) } & ( \nu , j)_{(x)}^{(2 ) } & ( \nu , j)_{(x)}^{(3 ) } \\[1 mm ] \hline & & & \\ 6 & \begin{array}{r|c|l } \cline{2 - 2 } 1 & & 1 \\ \cline{2 - 2 } \end{array } & \begin{array}{r|c|l } \cline{2 - 2 } 0 & & 0 \\ \cline{2 - 2 } \end{array } & \\ & & & \\
7 & \begin{array}{r|c|c|l } \cline{2 - 3 } 1 & & & 1 \\ \cline{2 - 3 } \end{array } & \begin{array}{r|c|l } \cline{2 - 2 } 0 & & 0
\\ \cline{2 - 2 } \end{array } & \\ & & & \\ 9 & \begin{array}{r|c|c|l } \cline{2 - 3 } 1 & & & 1 \\ \cline{2 - 3 } \end{array } & \begin{array}{r|c|l } \cline{2 - 2 } 0 & & 0 \\ \cline{2 - 2 } \end{array } & \\ & & & \\ 10 & \begin{array}{r|c|c|l } \cline{2 - 3 } 2 & & & 1 \\ \cline{2 - 3 } \end{array } & \begin{array}{r|c|l } \cline{2 - 2 } 0 & & 0 \\ \cline{2 - 2 } 0 & & 0 \\ \cline{2 - 2 } \end{array } & \begin{array}{r|c|l } \cline{2 - 2 } 0 & & 0
\\ \cline{2 - 2 } \end{array } \\ & & & \\ 11 & \begin{array}{r|c|c|l } \cline{2 - 3 } 3 & & & 1 \\ \cline{2 - 3 } \end{array } & \begin{array}{r|c|c|l } \cline{2 - 3 } 0&&&0\\ \cline{2 - 3 } 0&&\multicolumn{2}{l}{0}\\ \cline{2 - 2 } \end{array } & \begin{array}{r|c|l } \cline{2 - 2 } 0 & & 0 \\ \cline{2 - 2 } \end{array } \\ & & & \\ 13 & \begin{array}{r|c|c|l } \cline{2 - 3 } 1 & & & 1 \\ \cline{2 - 3 } \end{array } & \begin{array}{r|c|c|l } \cline{2 - 3 }
0&&&0\\ \cline{2 - 3 } 0&&\multicolumn{2}{l}{0}\\ \cline{2 - 2 } \end{array } & \begin{array}{r|c|l } \cline{2 - 2 } 0 & & 0 \\ \cline{2 - 2 } \end{array } \\ & & & \\ \hline \end{array}\ ] ] let @xmath768 be the involution on rigged configurations that complements the quantum numbers .
more precisely @xmath769 where @xmath770 is obtained from @xmath86 by complementing each partition @xmath699 within the box of width @xmath688 and height @xmath700 . by definition ,
the coquantum number bijection is @xmath771 .
it was shown in @xcite that the bijection @xmath772 preserves the statistics .
* theorem 8.3 ) [ thm statistics ] for @xmath633 we have @xmath773 proposition [ path lr ] yields a bijection @xmath774 .
an element @xmath775 will be mapped to @xmath776 where @xmath777 is the unique yamanouchi tableau of shape and content @xmath92 ( see the end of section [ sec : crystal a ] ) .
hence the above bijection restricts to a bijection between paths and lr tableaux @xmath778 .
this in turn induces a bijection between paths and rigged configurations .
it will be useful for later to state this bijection explicitly . by abuse of notation
we denote the bijection induced by @xmath727 also by @xmath779 , and similarly for @xmath728 .
let @xmath780 . recall that @xmath781 in @xmath588 has @xmath782 rows and @xmath783 columns .
factor each step @xmath784 into columns @xmath785 . under the bijection @xmath786 , which maps @xmath787 to @xmath788 , a letter @xmath789
is mapped to the letter @xmath790 in column @xmath120 in @xmath369 .
as before the rigged configuration @xmath703 corresponding to @xmath54 is constructed recursively , this time on @xmath791 , @xmath792 and increasing @xmath789 in this order .
suppose the letter @xmath120 is at height @xmath742 in @xmath793 .
then select singular strings of length @xmath794 in the @xmath795-th , @xmath796-th , ... , @xmath742-th rigged partition such that the @xmath743 are maximal subject to the condition @xmath797 .
as before add a box to each selected string , keeping it singular and leaving all other strings unchanged .
if @xmath789 and @xmath798 the vacancy number is calculated with respect to the sequence of rectangles @xmath799 where @xmath742 is the height of @xmath120 in @xmath784 .
let @xmath800 , @xmath801 and @xmath802 in @xmath803 .
then the corresponding lr tableau in @xmath804 is the @xmath369 of example [ ex : bij ] .
for example the letter @xmath805 in @xmath806 corresponds to the letter @xmath807 in the fourth column of @xmath369 and will select singular string lengths @xmath808 in the second and third rigged partition . similarly , the letter @xmath809 in @xmath810 corresponds to the letter @xmath811 in the third column of @xmath369 and will select a singular string of length @xmath812 in the second rigged partition .
for example , the rigged configuration in table [ table : ex ] for @xmath813 corresponds to the intermediary paths @xmath814 [ lem ineq ] let @xmath550 where @xmath815 is a single row with at least two boxes .
fix @xmath624 and let @xmath816 ( resp .
@xmath817 ) be obtained from @xmath369 by removing the rightmost ( resp .
two rightmost ) letters @xmath220 from @xmath369 . denote by @xmath743 ( resp .
@xmath818 ) the lengths of the selected strings under the bijection algorithm for @xmath727 going from @xmath816 to @xmath369 ( resp .
@xmath817 to @xmath816 ) .
then @xmath819 let @xmath820 , @xmath821 , and @xmath822 .
the vacancy numbers change by @xmath823 let @xmath120 ( resp .
@xmath824 ) be the column index of the box @xmath825 ( resp .
@xmath826 ) . then @xmath827 . since @xmath828 it follows from that there are no singular strings of length @xmath67 with @xmath829 in @xmath830 .
hence @xmath831 .
by induction on @xmath180 it follows that @xmath832 for @xmath833 . since by the strings in the range @xmath834
are nonsingular we must have @xmath835 .
* theorem 8.3 ) [ rc embed ] the following diagram commutes : @xmath836 [ cdual rc ] the following diagram commutes : @xmath837 the diagram also commutes with @xmath727 replacing @xmath728 . by corollary [ embedding dual ] and theorem [ rc embed ] , one may reduce to the case that @xmath550 is a sequence of single - rowed shapes .
recall that @xmath838 . by proposition
[ rc dual ] it is obvious that @xmath768 commutes with the duality map on rigged configurations . thus it suffices to prove the theorem where @xmath727 replaces @xmath728 .
we proceed by induction on @xmath220 .
the theorem is true for @xmath839 .
let @xmath624 and @xmath840 be the column indices of the letters @xmath220 in @xmath369 .
then in @xmath625 the column indices of the letters @xmath220 are in the alphabet @xmath841 with letters @xmath842 omitted . call them @xmath843 for @xmath844 .
let @xmath816 be the tableau obtained from @xmath369 by removing all letters @xmath220 and @xmath845 . by induction @xmath846 .
hence it suffices to show that the addition of the letters @xmath220 to @xmath816 under @xmath752 is ( up to reversal of the order of all partitions ) equal to the addition of the letters @xmath220 to @xmath847 under @xmath848 .
set @xmath849 and let @xmath850 be the rigged configuration corresponding to @xmath816 with letters @xmath220 added in columns @xmath851 . adding the letter @xmath220 in column @xmath852 has the effect on the rigged configurations of selecting singular strings in @xmath853 for @xmath854 of length @xmath855 maximal such that @xmath856 , adding a box to the selected strings and making them singular again . by lemma [ lem ineq ] , @xmath857 .
it follows that @xmath855 with @xmath858 and @xmath859 is uniquely defined to be the length of the maximal singular string in @xmath860 such that @xmath861 where @xmath862 if @xmath863 and @xmath864 if @xmath865 or @xmath866 .
define a matrix @xmath27 with @xmath153 rows and @xmath867 columns with entries @xmath868 .
the entries are weakly increasing along a row and weakly decreasing along a column .
some of the entries may be @xmath869 .
now do the analogous construction for @xmath625 . call the selected strings under @xmath727 , @xmath870 for @xmath844 and @xmath871 . by the same arguments as before they are uniquely defined as the lengths of the maximal singular string in @xmath872 such that @xmath873 where @xmath874 if @xmath875 and @xmath876 if @xmath865 or @xmath877 .
this yields a matrix @xmath878 with entries @xmath879 .
it is clear from that the entries in @xmath27 can either be defined inductively row by row , top to bottom , right to left , or column by column , right to left , top to bottom .
the same is true for @xmath878 .
since by induction @xmath880 for @xmath881 it follows that @xmath882 or equivalently @xmath883 .
this implies @xmath884 as desired .
let @xmath885 and @xmath886 then @xmath887 , @xmath888 , @xmath889 , @xmath890 and @xmath891 so that @xmath892 on the other hand @xmath893 , @xmath894 so that @xmath895 and @xmath896 so indeed @xmath897 . for @xmath898
there is a unique embedding of multiplicity - free @xmath899-crystals @xmath900 explicitly it is given by @xmath901 where @xmath902 . by , @xmath903 this map also preserves @xmath242 and @xmath405 when such operators act on the left tensor factor .
an analogous map can be defined on rigged configurations .
let @xmath588 be a sequence of rectangles such that @xmath904 and @xmath905 and @xmath906 the same sequence of rectangles with @xmath907 replaced by @xmath908 and @xmath815 replaced by @xmath909 .
define @xmath910 by the following commutative diagram : @xmath911 here @xmath912 is the rc - transpose map as defined in ( * ? ? ?
* section 7 ) .
it can be checked by direct computation that for @xmath913 the rigged configuration @xmath914 is obtained by adding a singular string of length @xmath395 to the @xmath180-th rigged partition for all @xmath915 .
[ thm : emb ] let @xmath45 be any tensor product of type @xmath36 crystals of the form @xmath916 and let @xmath37 ( resp .
@xmath906 ) be the sequence of rectangles associated with @xmath917 ( resp .
@xmath918 ) .
the following diagram commutes : @xmath919 the proof follows from the evacuation theorem ( * ? ? ?
* theorem 5.6 ) , the transpose theorem ( * ? ? ? * theorem 7.1 ) and @xmath920 for @xmath921 which is true by proposition [ pp : flip p ] .
let @xmath8 be an affine lie algebra of type @xmath21 , @xmath1 , @xmath22 , or @xmath2 , defined by the dynkin diagrams with distinguished vertex @xmath23 given in table [ tab : dynkin ] .
rl @xmath923 : + @xmath924 & ( 126,20)(-5,-5 ) ( 0,0)(20,0)3 ( 100,0)(20,0)2 ( 23,0)(20,0)2(1,0)14 ( 83,0)(1,0)14 ( 2.85,-1)(0,2)2(1,0)14.3 ( 102.85,-1)(0,2)2(1,0)14.3 ( 59,0)(4,0)6(1,0)2 ( 10,0)(0,0)@xmath925 ( 110,0)(0,0)@xmath926 ( 0,-5)(0,0)[t]@xmath23 ( 20,-5)(0,0)[t]@xmath927 ( 40,-5)(0,0)[t]@xmath928 ( 100,-5)(0,0)[t]@xmath929 ( 120,-5)(0,0)[t]@xmath399 + & + @xmath930 : & ( 26,20)(-5,-5 ) ( 0,0)(20,0)2 ( 2.958,-0.5)(0,1)2(1,0)14.084 ( 2.598,-1.5)(0,3)2(1,0)14.804 ( 0,-5)(0,0)[t]@xmath23 ( 20,-5)(0,0)[t]@xmath927 ( 10,0)(0,0)@xmath925 + & + @xmath931 : + @xmath924 & ( 126,20)(-5,-5 ) ( 0,0)(20,0)3 ( 100,0)(20,0)2 ( 23,0)(20,0)2(1,0)14 ( 83,0)(1,0)14 ( 2.85,-1)(0,2)2(1,0)14.3 ( 102.85,-1)(0,2)2(1,0)14.3 ( 59,0)(4,0)6(1,0)2 ( 10,0)(0,0)@xmath925 ( 110,0)(0,0)@xmath925 ( 0,-5)(0,0)[t]@xmath23 ( 20,-5)(0,0)[t]@xmath927 ( 40,-5)(0,0)[t]@xmath928 ( 100,-5)(0,0)[t]@xmath929 ( 120,-5)(0,0)[t]@xmath399 + & + @xmath932 : & ( 26,20)(-5,-5 ) ( 0,0)(20,0)2 ( 2.958,-0.5)(0,1)2(1,0)14.084 ( 2.598,-1.5)(0,3)2(1,0)14.804 ( 0,-5)(0,0)[t]@xmath23 ( 20,-5)(0,0)[t]@xmath927 ( 10,0)(0,0)@xmath926 + & + @xmath933 : + @xmath924 & ( 126,20)(-5,-5 ) ( 0,0)(20,0)3 ( 100,0)(20,0)2 ( 23,0)(20,0)2(1,0)14 ( 83,0)(1,0)14 ( 2.85,-1)(0,2)2(1,0)14.3 ( 102.85,-1)(0,2)2(1,0)14.3 ( 59,0)(4,0)6(1,0)2 ( 10,0)(0,0)@xmath926 ( 110,0)(0,0)@xmath926 ( 0,-5)(0,0)[t]@xmath23 ( 20,-5)(0,0)[t]@xmath927 ( 40,-5)(0,0)[t]@xmath928 ( 100,-5)(0,0)[t]@xmath929 ( 120,-5)(0,0)[t]@xmath399 + & + @xmath40 : + @xmath924 & ( 126,20)(-5,-5 ) ( 0,0)(20,0)3 ( 100,0)(20,0)2 ( 23,0)(20,0)2(1,0)14 ( 83,0)(1,0)14 ( 2.85,-1)(0,2)2(1,0)14.3 ( 102.85,-1)(0,2)2(1,0)14.3 ( 59,0)(4,0)6(1,0)2 ( 10,0)(0,0)@xmath926 ( 110,0)(0,0)@xmath925 ( 0,-5)(0,0)[t]@xmath23 ( 20,-5)(0,0)[t]@xmath927 ( 40,-5)(0,0)[t]@xmath928 ( 100,-5)(0,0)[t]@xmath929 ( 120,-5)(0,0)[t]@xmath399 there is an embedding @xmath934 of the weight lattice of @xmath8 into that of @xmath935 which preserves distance up to a fixed scalar factor .
this suggests that @xmath11-crystals can be embedded into @xmath936-crystals .
we introduce the notion of a virtual @xmath11-crystal , which is by definition a suitable subset of a @xmath936-crystal .
we define the virtual @xmath11-crystal @xmath20 and give some evidence that it agrees with the crystal @xmath17 of conjecture [ conj : brs ] ( whose existence was conjectured in ( * ? ? ?
* conj . 2.1 ) ) .
these conjectures are proved for crystals @xmath19 of types @xmath937 .
we use the notation of section [ subsec : affine case ] .
the null root @xmath938 is given explicitly by @xmath939 to each @xmath8 we associate two simple lie algebras @xmath32 and @xmath940 [ cols="^,^,^,^,^",options="header " , ] note that @xmath941 except for @xmath1 .
the algebra @xmath940 will be used in section [ sec : fermi ] .
let @xmath942 be the standard bilinear form on @xmath41 normalized by @xmath943 let @xmath944 we use the superscript and subscript @xmath371 to denote the root system @xmath935 ( or @xmath5 ) .
there is an embedding of weight lattices @xmath945 defined by @xmath946 this induces the embedding @xmath947 given by @xmath948 one may verify that @xmath949 observe that @xmath950 under the normalization .
note that if @xmath951 , then @xmath952 if and only if the following conditions hold : @xmath953 the third condition is a consequence of the other two .
let @xmath464 be the longest element in the weyl group of type @xmath5 .
then for @xmath954 , @xmath955 since this holds for @xmath956 for @xmath957 .
we wish to define a category of @xmath11-crystals which are realized as subsets of @xmath936-crystals .
let @xmath958 be a finite @xmath936-crystal .
in view of we define the virtual @xmath11 raising and lowering operators on @xmath958 by @xmath959 let @xmath960 be a nonempty subset such that @xmath961 then @xmath962 has a virtual weight function @xmath963 defined by @xmath964 this is well - defined since @xmath965 is injective .
a virtual @xmath11-crystal is a pair @xmath966 such that @xmath962 satisfies and is closed under the virtual raising and lowering operators @xmath93 and @xmath94 for @xmath967 . sometimes the larger crystal @xmath958 ( called the ambient crystal ) is omitted in the notation .
[ pp : virtual crystal ] a virtual @xmath11-crystal @xmath966 is a @xmath11-crystal in the sense of section [ subsec : crystals ] ( see also [ subsec : affine case ] )
. all the properties are immediate except for . since @xmath93 and @xmath968 ( resp . @xmath94 and @xmath969 ) commute for @xmath970 , by the definitions we have for @xmath971 @xmath972 here @xmath973 denotes the largest integer smaller or equal to @xmath741 .
first let us verify for @xmath490 .
we have that @xmath974 is a multiple of @xmath975 by applied to @xmath976 and for @xmath54 viewed as an element of @xmath958 .
for @xmath977 we have @xmath978 by , , and .
a similar calculation establishes for @xmath979 .
let @xmath970 .
observe that @xmath980 using for @xmath981 and .
suppose first that @xmath982
. then @xmath983 and one has @xmath984 if @xmath985 then @xmath986 and @xmath987 in either case follows by and .
a morphism of virtual @xmath11-crystals @xmath988 is a morphism @xmath989 of the ambient @xmath936-crystals which restricts to a morphism @xmath990 of @xmath11-crystals .
let @xmath991 and @xmath992 be virtual @xmath11-crystals .
we wish to define their tensor product .
define the virtual crystal operators @xmath93 and @xmath94 on @xmath993 by .
define @xmath994 by .
unfortunately the subset @xmath995 need not be closed under @xmath93 and @xmath94 .
let @xmath996 , @xmath997 , and let @xmath998 be @xmath936-crystals .
let @xmath992 and @xmath991 be the virtual crystals generated by the elements @xmath517 and @xmath515 where @xmath999 then @xmath1000 .
this disagrees with the usual tensor product structure on @xmath11-crystals defined by .
let @xmath981 .
consider the following conditions .
@xmath1001 say that @xmath54 is @xmath67-aligned for @xmath970 if the first condition holds , @xmath23-aligned if the second holds , and @xmath399-aligned if the third holds .
say that @xmath966 is an aligned virtual @xmath11-crystal if , for every @xmath977 , @xmath54 is @xmath67-aligned for all @xmath1002 .
let @xmath970 .
if @xmath54 is @xmath67-aligned then @xmath1003 if @xmath54 is @xmath399-aligned then @xmath1004 if @xmath54 is @xmath23-aligned then @xmath1005 [ rem : aligned ] let @xmath966 be a virtual @xmath7-crystal and @xmath981 . if @xmath1006 then @xmath54 is @xmath399-aligned and if @xmath1007 then @xmath54 is @xmath23-aligned
. [ pp : virtual tensor ] aligned virtual @xmath11-crystals form a tensor category .
let @xmath1008 be aligned virtual @xmath11-crystals for @xmath266 .
since they are aligned , it follows that for all @xmath1009 and all @xmath48 , @xmath93 and @xmath94 either act entirely on the left factor or on the right factor of @xmath1010 .
therefore the operators @xmath93 and @xmath94 coincide with those given by the @xmath11-crystal structure on the tensor product @xmath1011 defined by and , which define a tensor category . finally , it is easy to verify that @xmath1012 is aligned .
if @xmath992 and @xmath991 are aligned virtual @xmath11-crystals then define their tensor product to be the aligned virtual @xmath11-crystal @xmath1013 .
say that the virtual @xmath11-crystal @xmath966 is simple if 1 .
@xmath958 is a simple @xmath936-crystal .
2 . @xmath962 , with its virtual @xmath11-crystal structure , is isomorphic to a simple @xmath11-crystal .
3 . @xmath1014 .
the following result shows that local isomorphisms and energy functions exist for aligned simple virtual crystals .
[ pp : virtual r ] let @xmath991 and @xmath992 be aligned simple virtual @xmath11-crystals .
write @xmath1015 for the local isomorphisms and energy functions for the pair of simple @xmath936-crystals @xmath1016 and @xmath1017 , and the pair of simple @xmath11-crystals @xmath1018 and @xmath1019
. then @xmath1020 @xmath1012 is an aligned simple virtual @xmath11-crystal , by proposition [ pp : virtual tensor ] and theorem [ thm : simple ] , generated from the extremal vector @xmath1021 which holds by and property 3 of the definition of simple virtual crystal @xmath1008 .
similarly @xmath1022 is an aligned simple virtual @xmath11-crystal with generator @xmath1023 .
now @xmath1024 is a @xmath936-crystal isomorphism such that @xmath1025 . as
such it intertwines with the virtual operators @xmath93 and @xmath94 for @xmath1002 .
therefore @xmath1026 is a @xmath11-crystal isomorphism from the @xmath11-component of @xmath362 to that of @xmath453 , that is , from @xmath1011 to @xmath1027 .
but there is a unique such map , namely , @xmath168 .
this proves . by abuse of notation ,
let @xmath1028 be defined by .
it suffices to show that @xmath163 satisfies the defining properties of the energy function given in theorem [ thm : two tensor ] .
this is entirely straightforward except for .
let @xmath1029 for @xmath266 and write @xmath166 .
suppose first that @xmath1030 and @xmath1031 .
since @xmath1008 is aligned for @xmath266 , this means @xmath1032 and @xmath1033 by .
applying @xmath1034 to @xmath1010 @xmath975 times and using , one has @xmath1035 dividing by @xmath975 one obtains for @xmath163 in the first case .
the other cases are similar .
let @xmath966 be an aligned simple virtual @xmath11-crystal .
suppose the ambient simple crystal @xmath958 is graded by the intrinsic energy function @xmath1036 .
then @xmath962 has an inherited intrinsic energy function @xmath1037 , defined by @xmath1038 call @xmath1039 an aligned graded simple virtual crystal .
let @xmath1040 be an aligned graded simple virtual @xmath11-crystal for @xmath1041 .
consider the simple @xmath11-crystal @xmath1042 and the simple @xmath936-crystal @xmath1043 .
there are two ways to form an intrinsic energy function for @xmath962 .
one way is to use the fact that it is a tensor product of graded simple @xmath11-crystals @xmath1044 using .
the other way is to use the inherited intrinsic energy function coming from the fact that @xmath966 is an aligned virtual crystal and @xmath1045 is a graded simple crystal ( with @xmath1036 defined by in terms of @xmath1046 ) .
it follows from that the two definitions agree . for @xmath8 of type @xmath21 , @xmath1 , @xmath22 , or @xmath2
, we wish to define a virtual @xmath11-crystal @xmath20 ( @xmath14 ) .
let @xmath1047 be the @xmath936-crystal @xmath1048 the virtual crystal @xmath1049 is defined as the subgraph of @xmath1047 generated from the extremal vector @xmath1050 by applying the virtual operators @xmath93 and @xmath94 ( @xmath1002 ) .
[ conj : aligned ] @xmath20 is aligned . for @xmath957 ,
define @xmath1051 the crystals @xmath17 of conjecture [ conj : brs ] come equipped with the following prescribed decomposition as @xmath31-crystals : @xmath1052 where @xmath1053 has the form @xmath1054 such that @xmath1055 , @xmath1056 , and @xmath1057 for @xmath1058
. moreover , @xmath17 comes equipped with the following intrinsic energy function : @xmath1059 [ conj : virtual ] the virtual crystal @xmath20 is the simple crystal @xmath17 of the @xmath11-module @xmath311 , with extremal vector @xmath1060 of weight @xmath1061 . in particular , @xmath20 has the @xmath31-crystal decomposition given by and @xmath1062
. it will be shown in proposition [ pp : big enough ] that @xmath20 has at least the @xmath31-components specified by and that the virtual intrinsic energy on these components agrees with . if either @xmath0 or @xmath18 , it will be shown in theorems [ thm : d decomp ] , [ thm : c decomp ] , [ thm : a2 decomp ] , and [ thm : a2d decomp ] that @xmath20 is aligned and has exactly the decomposition .
we now consider the problem of giving explicit conditions for membership in the virtual crystal @xmath20 as a subset of @xmath1047 .
we shall see that a necessary condition for membership in @xmath20 is self - duality up to local isomorphism .
let @xmath1063 be the @xmath936-crystal obtained by reversing the order of the tensor factors in @xmath1047 .
let @xmath1064 be defined as @xmath20 is , except with @xmath1047 replaced by @xmath1063 .
there is a @xmath936-crystal isomorphism @xmath1065 where @xmath168 is the identity or the local isomorphism according as @xmath1047 has one or two tensor factors . the map @xmath168 induces a @xmath11-crystal isomorphism @xmath1066 .
[ pp : brs self dual ] for every @xmath1067 , @xmath1068 first let @xmath1069 .
since @xmath1063 is multiplicity - free as a @xmath899-crystal , it suffices to show that both @xmath1070 and @xmath1071 are @xmath5-highest weigh vectors of the same weight . since @xmath168 is a @xmath936-crystal isomorphism , @xmath1071 is a @xmath899-highest weight vector of weight @xmath976 .
@xmath1072 is an @xmath5-highest weight vector of weight @xmath1073 by . since @xmath1074 it follows that @xmath1075 by .
so holds .
now suppose @xmath1076 satisfies and @xmath1077 for some @xmath1002 . it will be shown that @xmath1078 also satisfies .
this , together with a similar proof with @xmath93 replacing @xmath94 , suffices .
let @xmath1079 .
we have @xmath1080 by , theorem [ thm : flip ] , , and the fact that @xmath168 is a @xmath936-crystal isomorphism .
holds for @xmath1081 and @xmath1082 by applying with @xmath1083 and @xmath1084 respectively . holds for @xmath1078 for @xmath970 by observing that @xmath1085 and @xmath1086 commute , and applying for @xmath1087 and then for @xmath1088 .
[ pp : virtual eps ] for all @xmath1076 such that holds and for all @xmath1089 , @xmath54 is @xmath67-aligned .
let @xmath970 .
since @xmath168 is an isomorphism of @xmath936-crystals we have @xmath1090 by theorem [ thm : flip ] and .
the proof for @xmath156 is similar .
recall that for the affine algebras @xmath8 under consideration @xmath32 is either of type @xmath3 or @xmath4 .
a pair @xmath966 is a virtual @xmath31-crystal provided that @xmath958 is a finite @xmath899-crystal , @xmath1091 is a subset that satisfies and is closed under the virtual operators @xmath93 and @xmath94 for @xmath957 defined by .
an aligned virtual @xmath31-crystal is one that consists of elements that are @xmath67-aligned for @xmath1092 .
let @xmath1053 and @xmath1093 .
since the virtual operators @xmath93 and @xmath94 ( for @xmath1092 ) are comprised of type @xmath899 operators and the latter preserve the property of being a tableau of a given shape , it follows that @xmath1094 is closed under these virtual operators
. therefore we may define @xmath39 to be the virtual @xmath31-crystal given by the subset of @xmath1094 generated from @xmath1095 using the virtual operators @xmath93 and @xmath94 for @xmath957 .
the following crucial theorem is due to t. baker @xcite .
our approach , which emphasizes the self - duality property , could be used to give an elegant proof of this result .
[ thm : virtual fin ] @xmath39 is aligned and is isomorphic to @xmath91 as a @xmath31-crystal .
[ pp : align hwv ] let @xmath966 be an aligned @xmath31-crystal . then every @xmath977 is in the connected component of a virtual @xmath31-highest weight vector @xmath362 , and @xmath362 is a @xmath899-highest weight vector when viewed as an element of @xmath958 . let @xmath977 .
suppose @xmath1096 for some @xmath1097 .
since @xmath966 is an aligned virtual @xmath31-crystal , it follows that @xmath1098 where @xmath1099 for @xmath1100 and @xmath1101 for @xmath1102 . replacing @xmath54 with @xmath1103 and continuing in this manner
, eventually one has @xmath1104 for @xmath1097 .
[ pp : virtual b ] let @xmath140 and @xmath1105 . if @xmath1106 then @xmath1107 moreover if @xmath1108 and @xmath54 satisfies then @xmath1106 .
the proof that elements of @xmath39 satisfy is similar to the proof of proposition [ pp : brs self dual ] and is omitted .
suppose @xmath1108 .
let @xmath1109 be such that holds .
let @xmath1110 be the virtual @xmath1111-crystal generated by @xmath54 .
it suffices to show that @xmath1112 . by the proof of proposition [ pp : brs self dual ] , holds for any element in @xmath1113 . but implies that @xmath1113 is @xmath67-aligned for @xmath970 , in a manner similar to the proof of proposition [ pp : virtual eps ] . by remark [ rem : aligned ] @xmath1113 is @xmath399-aligned .
therefore @xmath1114 is an aligned @xmath1111-crystal . by proposition [ pp : align hwv ] @xmath1112 . for this section let @xmath1116 .
the crystal graphs of the irreducible integrable @xmath1115-modules @xmath91 for @xmath140 , were constructed explicitly by kashiwara and nakashima @xcite in terms of certain tableaux . the crystal graph @xmath354 is given by @xmath1117 with @xmath1118 for @xmath957 .
let @xmath1119 be the set of vertices of @xmath354 , totally ordered so that the elements appear from smallest to largest going from left to right in .
let @xmath1120 denote the length of the word @xmath362 .
a subset @xmath1121 $ ] is often identified with the column word having precisely the letters in @xmath371 .
write @xmath1122-a$ ] , @xmath1123 , and @xmath1124}}-{\overline{a}}={\overline{a^c}}$ ] .
in particular if @xmath362 and @xmath608 are column words in the alphabet @xmath353 $ ] then @xmath1125 is a column word in the alphabet @xmath1119 and @xmath1126 .
if @xmath1127 is a column word in the alphabet @xmath1119 then it can be written uniquely in the form @xmath1128 where @xmath362 and @xmath608 are column words in the alphabet @xmath353 $ ] . in this case
write @xmath1129 and @xmath1130 .
the crystal @xmath1131 may be defined as the connected component in the tensor product @xmath1132 of the column word @xmath1133 where the tensor symbols are omitted .
explicitly , @xmath1131 is the set of column words @xmath41 of length @xmath411 in the alphabet @xmath353\cup\overline{[n]}$ ] that satisfy the one - column condition @xcite @xmath1134}}\right| } + { \left|{{p_-}|_{[i]}}\right| } \le i$.}\ ] ] let @xmath1053 .
write @xmath1135 where @xmath1136
. then @xmath91 may be defined as the connected component of @xmath1137 in @xmath1138 .
[ lem:1cc tableau ] @xmath41 satisfies if and only if @xmath1139 is a tableau .
the following are equivalent : 1 .
@xmath41 does not satisfy .
2 . there is an index @xmath957 such that @xmath1140 , @xmath1141 , and @xmath1142}}\right| } > { \left|{{p^c_-}|_{[i]}}\right|}$ ] .
3 . there is an index @xmath957 such that @xmath1142}}\right| } > { \left|{{p^c_-}|_{[i]}}\right|}$ ] .
@xmath1139 is not a tableau .
each of the above conditions is obviously equivalent to the next except for 3 implies 2 .
for that case just take @xmath67 to be minimal . if @xmath1143 or @xmath1144 then the index @xmath1145 satisfies 3 , contradicting minimality .
we now recall the explicit isomorphism @xmath1146 given in @xcite .
the first case is @xmath1147 . for @xmath1148 ,
let @xmath1149 where the maximum is computed with respect to the partial order on column words that says @xmath1150 if and only if @xmath1151 is a tableau .
this given , let @xmath1152 define the map @xmath1153 by @xmath1154 where @xmath1155 is regarded as the @xmath899-subcrystal of @xmath1156 by taking the column - reading word .
lemma [ lem:1cc tableau ] ensures that the column - reading word indeed corresponds to a tableau .
let @xmath1157 for @xmath1158 , @xmath1159 , and @xmath1160 where @xmath1161 .
define @xmath1162 by @xmath1163 we observe that @xmath1164 and @xmath1165 may be computed from @xmath41 by local isomorphisms of type @xmath371 for tensor products of columns .
[ lem : one col iso ] @xmath1166 the construction in the definition of @xmath1164 and @xmath1165 is equivalent to the algorithm of section [ sec : r sc ] used to compute @xmath1167 .
let @xmath1168 and @xmath1169 .
then @xmath1170 , @xmath1171 , @xmath1172 , @xmath1173 , @xmath1174 , and @xmath1175 . on the other hand @xmath1167
may be computed using the algorithm described in section [ sec : r sc ] .
note that @xmath1176 so that @xmath1177 @xmath1178 is the set of all heights which has a dot on the right , but no dot on the left .
@xmath1179 is the set of heights with a dot on the left , but no dot on the right .
hence @xmath86 is obtained by selecting from top to bottom for each element @xmath1180 the maximal @xmath1181 with a dot on the left and no dot on the right at height @xmath1182 .
the elements in @xmath86 are all selected by the algorithm for the computation of @xmath168 as described in section [ sec : r sc ] , and hence both computations of @xmath86 are equivalent . for the following lemma , we identify @xmath1119 with @xmath1183 $ ] via the bijection @xmath1184 and @xmath1185 for @xmath360 $ ] .
[ lem : self dual ] let @xmath362 and @xmath608 be column words in the alphabet @xmath353 $ ] . then there is a unique pair of column words @xmath453 and @xmath1186 in the alphabet @xmath353 $ ] such that @xmath1187 , @xmath1188 , and ( defining @xmath1189 and @xmath1190 ) @xmath1191 where @xmath1192 is the local isomorphism of type @xmath935 with @xmath1193 . for existence ,
define @xmath453 and @xmath1186 by @xmath1194 by definition @xmath453 and @xmath1186 are column words of the correct length . applying the map @xmath1195 to , by propositions [ pp : r dual ] and [ pp : r ev ] and
the fact that @xmath168 is an involution , one has @xmath1196 by definition @xmath1189 and @xmath1190 , so that @xmath395 and @xmath1197 are column words of the same length and @xmath369 and @xmath1198 are column words of the same length . thus to prove it is enough to show that @xmath1199 .
since the shapes of both @xmath1200 and @xmath1201 have two columns , it suffices to show that 1 .
@xmath1202 for @xmath1203 $ ] and for @xmath1204}}$ ] .
@xmath1200 and @xmath1201 have the same shape . now @xmath1205 and @xmath1206 .
we have @xmath1207})={\mathbb{p}}(u'v')={\mathbb{p}}(vu)={\mathbb{p}}(st|_{[n]})\ ] ] by and @xmath1208 } } } ) = { \mathbb{p}}({\overline{v}}^c { \overline{u}}^c ) = { \mathbb{p}}({\overline{u'}}^c { \overline{v'}}^c ) = { \mathbb{p}}(st|_{{\overline{[n]}}})\ ] ] by .
this proves 1 .
condition 2 is equivalent to : @xmath1209 and @xmath1210 have the same shape . but
@xmath1211 by proposition [ pp : record dual ] with @xmath1212 .
so it is enough to show that @xmath1213 . but
this holds since @xmath1214 has at most two rows and exactly @xmath1215 cells ( since @xmath596 is taken within the @xmath1216 rectangle ) . for uniqueness , observe that @xmath1217 is a tableau with @xmath1218 ones and @xmath1219 twos , and @xmath1220 is a tableau of the same shape with @xmath1219 ones and @xmath1218 twos .
the second tableau is uniquely specified by this property .
[ pp : virtual c ] let @xmath1221 .
write @xmath1222}$ ] and @xmath1223}$ ] and similarly for @xmath608
. the following are equivalent : 1 .
2 . holds , @xmath1225 is a tableau and @xmath1226 .
3 . holds , @xmath1227 is a tableau and @xmath1228 .
1 implies 2 : let @xmath1229 under the @xmath1115-crystal isomorphism @xmath1230 given by .
then @xmath1231 , @xmath1232 , @xmath1233 , and @xmath1234 with @xmath1235 defined by lemma [ lem : one col iso ] . by lemma [ lem:1cc tableau ] @xmath1236 is a tableau , with @xmath1237 .
2 implies 1 : let @xmath41 be such that @xmath1238}$ ] and @xmath1239}$ ] . by lemma [ lem:1cc tableau ] @xmath1148 .
by lemma [ lem : self dual ] it follows that @xmath1240 under the isomorphism @xmath1230 .
for the equivalence of 2 and 3 , suppose that @xmath41 and @xmath586 are such that holds
. then the following are equivalent : 1 .
@xmath1241 is a tableau .
@xmath1242 is the column - reading word of a skew two - column tableau with unique southeast corner .
3 . @xmath1243 is a tableau .
4 . @xmath1244 is a tableau .
moreover it is easily seen that @xmath1245 .
now suppose 2 holds .
write @xmath54 in the form @xmath1246 . by the proof of lemma [ lem : self dual ] it follows that holds .
thus 3 follows .
the proof that 3 implies 2 is entirely similar .
[ pp : big enough ] let @xmath1053 be such that @xmath91 is a summand specified by . then there is a unique @xmath31-crystal embedding @xmath1247 defined by @xmath1248 where @xmath1249 , and @xmath1062 on @xmath1250 .
let @xmath1251 be any weight such that @xmath1252 occurs in the @xmath899-crystal decomposition of @xmath1047 .
@xmath1253 occurs in @xmath1047 with multiplicity one , and on this component @xmath1254 has value @xmath23 if @xmath1047 has one tensor factor and value @xmath1255 otherwise @xcite .
let @xmath1256 be the unique @xmath899-highest weight vector in @xmath1047 with @xmath1257 .
there is an embedding of virtual @xmath31-crystals @xmath1258 given by @xmath1259 where @xmath1260 .
the map is a morphism of virtual @xmath31-crystals since @xmath486 is a morphism of @xmath899-crystals .
now let @xmath1053 be a weight appearing in .
then @xmath1253 occurs in @xmath1047 . by the argument in the previous paragraph it suffices to show that @xmath1261 .
it can be shown that there is an explicit sequence of virtual operators going from @xmath1050 to @xmath1262 .
moreover one has @xmath1263 if @xmath1047 has a single tensor factor , and otherwise @xmath1264 instead of writing down this sequence of operators in full generality , we give the following illustrative example
. let @xmath1265 , @xmath1266 , @xmath414 , @xmath1267 , and @xmath1268 .
we work in @xmath1064 and @xmath1063 for convenience . using the explicit rules for computing the operators of type @xmath935
, @xmath1269 each of the above vectors has the form @xmath1270 , where @xmath54 is given respectively by @xmath1271 this shows that @xmath1261 .
we characterize the elements of @xmath20 inside @xmath1047 using the self - duality condition .
[ pp : virtual d ] let @xmath0 and @xmath1076
. then @xmath1272 if and only if holds .
the forward direction holds by proposition [ pp : brs self dual ] . for the reverse direction , suppose @xmath1273 satisfies .
let @xmath1274 be the virtual @xmath11-crystal generated by @xmath54 .
it follows from the proof of proposition [ pp : brs self dual ] that every element of @xmath962 satisfies . but @xmath962 is automatically aligned by remark [ rem : aligned ] and proposition [ pp : virtual eps ] . by proposition [ pp : align hwv ] there is a @xmath31-highest weight vector @xmath1275 that is also a @xmath899-highest weight vector . since @xmath362 satisfies , @xmath1276 . since @xmath1277 it follows that @xmath1278 for some @xmath92 appearing in . by proposition [ pp :
big enough ] @xmath1279 .
it follows that @xmath1280 .
[ thm : d decomp ] let @xmath0 .
then the @xmath11-crystal @xmath20 is aligned and has the @xmath31-crystal decomposition given in .
this follows from the proof of proposition [ pp : virtual d ] .
[ lem : dual wt ] let @xmath54 be an element of the @xmath899-crystal @xmath1281 such that holds for @xmath1282 .
let @xmath368 be the number of occurrences of @xmath191 in @xmath54 for @xmath1283 .
then @xmath1284 for all @xmath1100 .
the lemma follows directly from @xmath1285 , which holds by for @xmath1282 , and .
[ thm : c decomp ] let @xmath1265 .
then @xmath25 is an aligned @xmath1286-crystal and @xmath1287 as a @xmath1115-crystal .
recall that the characterization of the @xmath1115-crystal @xmath1288 was given in proposition [ pp : virtual c ] . by proposition [ pp :
big enough ] @xmath1289 and the embedding is inclusion . to show equality it suffices to show that @xmath1290 is closed under @xmath405 and @xmath242 .
to check this we use the explicit computation of the @xmath23-string in @xmath936-crystals given in section [ sec : a perf ] .
let @xmath1291 such that @xmath1292 .
. then @xmath362 and @xmath608 both contain @xmath1215 and do not contain @xmath927 .
computing @xmath1294 on @xmath54 , in all cases @xmath1127 and @xmath741 are obtained from @xmath362 and @xmath608 respectively by removing @xmath1215 from the bottom and adding @xmath927 at the top .
thus @xmath1295 is a tableau .
@xmath1296 satisfies by proposition [ pp : brs self dual ] .
let @xmath1297 and @xmath1298 be defined as @xmath526 for @xmath266 .
then @xmath1299 is a tableau , being obtained from @xmath1300 by removing @xmath1215 from the bottom of each column . from this one
also sees that @xmath1301 . by proposition [ pp : virtual c ] , @xmath1302 . in an entirely similar manner
, one may show that @xmath1290 is closed under @xmath242 . by theorem [ thm : virtual fin ]
it remains to show that the elements of @xmath1303 are @xmath23-aligned . by lemma [ lem : dual wt ] , we need only check the case that @xmath927 and @xmath1215 occur in @xmath54 once each . since @xmath1151 is a tableau , @xmath1304 .
if @xmath1305 then @xmath1306 and @xmath54 is @xmath23-aligned .
so assume @xmath1307 .
if @xmath898 , then @xmath1308 has two tensor factors , and by and @xmath54 is @xmath23-aligned .
@xmath54 is @xmath23-aligned .
so let @xmath1309 .
write @xmath1310 and @xmath1311
. we will show that @xmath1312 , as the proof of @xmath1313 is similar .
@xmath1314 by .
thus it is enough to show that @xmath1315 . since @xmath1316 has a single @xmath1317 and a single @xmath1215 , this holds if @xmath1317 is in the right hand column of @xmath1316 . but
@xmath1318}= { \mathbb{p}}(b|_{[2,2n]})={\mathbb{p}}(u \widehat{v})={\sigma}(u\otimes \widehat{v})\ ] ] by .
so it suffices to show that @xmath1317 remains in the right hand column in passing from @xmath1319 to @xmath1320 .
but there is only one letter that moves from the right hand column to the left under @xmath168 , and it must be in @xmath1321 ( and therefore is not @xmath1317 ) since @xmath1322}$ ] is a tableau with columns of equal size .
[ pp : virtual a2 ] let @xmath1323 , @xmath1324 , @xmath1325 , and @xmath526 and @xmath546 defined as in proposition [ pp : virtual c ] for @xmath266 .
then @xmath971 if and only if 1 .
@xmath54 satisfies .
2 . @xmath1227 is a tableau .
3 . @xmath1326 .
the following are equivalent : * @xmath54 satisfies . * @xmath1327 .
* @xmath1282 satisfies .
the equivalence of the first two items follows from the fact that @xmath1308 is multiplicity - free as a @xmath899-crystal and @xmath486 is a @xmath899-crystal morphism .
the equivalence of the second and third items is a consequence of propositions [ pp : dual p ] and [ pp : flip p ] , and the definition of @xmath168 .
write @xmath1328 where @xmath453 and @xmath1186 are column words .
since @xmath362 and @xmath608 are column words of lengths @xmath1329 and @xmath411 respectively , it follows that @xmath453 and @xmath1186 have lengths @xmath1330 and @xmath180 for some @xmath1331 .
moreover one can pass between @xmath54 and @xmath1282 using a two column jeu de taquin .
let @xmath1332 and @xmath1333 be defined in a manner similar to @xmath526 for @xmath266 .
suppose first that @xmath977 .
then @xmath1334 for some @xmath1331 . by proposition [ pp : virtual c ] , @xmath1282 satisfies . by the above argument @xmath54 satisfies . by proposition [ pp : virtual
c ] @xmath1335 is a tableau and @xmath1336 .
to finish the forward direction it is enough to show that @xmath1337 and @xmath1338 . in passing from @xmath1339 to @xmath1340 ,
some letters go from the left column to the right .
since all letters in @xmath1341 are blocking letters of @xmath1342 , the only letters that can block @xmath1343 from moving to the right are those in @xmath1344 .
but by proposition [ pp : virtual c ] , @xmath1345 , so there are @xmath1346 letters in @xmath1343 available to move to the right
. the smallest @xmath1347 of these , actually do move .
in particular @xmath1337 and @xmath1338 . for the reverse direction
, suppose @xmath1348 satisfies the three properties . since @xmath1282 is a tableau , it suffices to show that @xmath1282 satisfies the properties in proposition [ pp : virtual c ] for @xmath180 . since @xmath54 satisfies , @xmath1282 satisfies , arguing as above .
this time , passing from @xmath1340 to @xmath1339 , we assume that @xmath1227 is a tableau .
that means that all the letters of @xmath1349 can not move to the left , and so stay in their column .
the left column only gets larger , and can only get larger by letters in the interval @xmath353 $ ] . therefore @xmath1337 and @xmath1338 again .
[ thm : a2 decomp ] let @xmath1323 .
then the @xmath1350-crystal @xmath25 is aligned , and as a @xmath1115-crystal , @xmath1351 moreover , the @xmath1115-crystal embedding @xmath1352 can be computed by @xmath1353 with @xmath1354 as defined in section [ subsec : embedding ] .
observe that alignedness follows from the above @xmath1115-crystal decomposition , since @xmath1007 for @xmath1323 and the virtual @xmath1115-crystals @xmath1355 are @xmath67-aligned for @xmath957 .
the map @xmath1356 coincides with @xmath1357 by proposition [ pp : big enough ] , since both are @xmath1115-crystal embeddings @xmath1358 .
let @xmath962 be as in proposition [ pp : big enough ] , which asserts that @xmath1359 . to show equality it suffices to show that @xmath962 is closed under @xmath242 and @xmath405 .
recall that for @xmath1323 , @xmath1360 and @xmath1361 .
let @xmath1362 be such that @xmath1292 .
write @xmath1363 .
let @xmath1297 and @xmath1298 be defined as @xmath526 is for @xmath266 .
it will be shown that @xmath1081 satisfies the three properties in proposition [ pp : virtual a2 ] . since @xmath1364
, @xmath1081 satisfies by proposition [ pp : brs self dual ] .
suppose first that @xmath1365 .
then @xmath1366 and @xmath1367 .
it is easily seen that properties 2 and 3 of proposition [ pp : virtual a2 ] hold for @xmath1081 , so that @xmath1368 .
otherwise suppose @xmath1369 .
then @xmath1370 , @xmath1305 , @xmath1371 , and @xmath1127 is obtained from @xmath362 by removing @xmath1215 from the bottom and putting @xmath927 at the top .
in particular @xmath1372 and @xmath1373 .
the only way that @xmath1374 fails to be a tableau is if @xmath1375 , when the first column is shorter than the second . since @xmath1227 is a tableau with columns of equal length , @xmath1376 . by lemma [ lem : dual wt ]
there are no ones present .
but then @xmath1377 , contrary to assumption .
therefore @xmath1299 is a tableau . to check property 3 of proposition [ pp : virtual a2 ] it
is enough to show @xmath1378 .
suppose not , that is , equality holds .
following the proof of proposition [ pp : virtual a2 ] , write @xmath1379 and define @xmath1332 and @xmath1333 . then @xmath1380 .
as in the aforementioned proof it can be shown that @xmath1381 for some @xmath1331 and that @xmath1337 and @xmath1338 . by our assumption
it follows that @xmath1382 and hence that @xmath1383 and @xmath1384 .
in particular @xmath1151 is a tableau . by lemma [ lem : dual wt ] @xmath1385 or @xmath1307 but not both . but
@xmath1385 is not possible since @xmath1151 is a tableau and @xmath1371 . and if @xmath1307 then @xmath1386 , contrary to our assumption .
the proof that @xmath962 is closed under @xmath242 , is similar .
[ pp : virtual a2d ] let @xmath1387 and @xmath1388 .
recall that @xmath1389 . then @xmath1390 if and only if @xmath1391 , @xmath54 is a tableau , and @xmath54 satisfies .
this follows from proposition [ pp : virtual b ] .
[ thm : a2d decomp ] let @xmath1387 .
then the virtual @xmath1392-crystal @xmath25 is aligned and is isomorphic to @xmath1393 as a @xmath1111-crystal .
it suffices to show that @xmath1394 is @xmath23-aligned and closed under @xmath1395 and @xmath1294 .
let @xmath1396 . by proposition [ pp : virtual a2d ] and lemma [ lem : dual wt ] , @xmath54 is a tableau and @xmath1397 where @xmath368 is the number of occurrences of @xmath191 in @xmath54 for @xmath1398 . write @xmath1362 where @xmath362 and @xmath608 are column words of type @xmath5 .
suppose first that @xmath887
. then @xmath1399 and @xmath1313 , so that @xmath54 is @xmath23-aligned .
writing @xmath1400 the column words @xmath1127 and @xmath741 are obtained from @xmath362 and @xmath608 by removing @xmath927 from the top and putting @xmath1215 at the bottom .
clearly @xmath1401 is a tableau .
since @xmath1402 it satisfies by proposition [ pp : brs self dual ] and is therefore in @xmath1393 by proposition [ pp : virtual a2d ] .
an entirely similar argument applies for the case @xmath1403 .
finally , suppose @xmath1404 . since @xmath54 is a tableau , @xmath1304 .
regardless of where the single symbol @xmath1215 appears in @xmath54 , one has @xmath1306 , so that @xmath54 is again @xmath23-aligned .
the goal of this section is to establish conjecture [ conj : virtual ] for @xmath25 of type @xmath1406 .
the @xmath1407-module @xmath1408 exists and has a simple crystal basis @xmath19 @xcite .
we use the symmetry of the dynkin diagram @xmath1406 to derive properties which uniquely define the structure of the affine crystal @xmath19 .
then we show that the virtual crystal @xmath25 satisfies these properties .
we state some generalities which must be satisfied by @xmath17 assuming it exists .
consider the automorphism of the dynkin diagram @xmath1406 given by @xmath1409 for @xmath1002 ; it has order two .
it induces an involution on crystal bases of @xmath1407-modules .
let @xmath1410 be the linear involution given by @xmath1411 for @xmath1412 .
there is an induced linear involution @xmath1413 given by @xmath1414 for all @xmath326 . identifying @xmath1415
one has @xmath1416 the existence of the dynkin diagram automorphism implies the following result .
[ pp : c brs ] let @xmath45 be the crystal basis of an irreducible integrable @xmath1286-module . then there is an automorphism @xmath1417 of @xmath45 with the following properties : @xmath1418 in particular @xmath1419 since the @xmath1115-crystal structure of @xmath17 is prescribed by , it only remains to determine the automorphism @xmath1417 .
[ rem : c auto ] let @xmath1420 . by proposition [ pp : c brs ] @xmath1417 acts on the set of @xmath1421-highest weight vectors in @xmath17 and is uniquely determined by this action . define the map @xmath1422 by @xmath1423 where @xmath396 is the order @xmath1215 automorphism of type @xmath1424 .
[ pp : virtual auto ] the map @xmath1417 defined by stabilizes the virtual crystal @xmath20 and satisfies the conditions of proposition [ pp : c brs ] for @xmath20 . @xmath1417 satisfies the conditions of proposition [ pp : c brs ] for all elements of @xmath1047 ; this follows immediately from the known properties of @xmath396 recorded in theorem [ thm : a perf ] . to see that @xmath1417 stabilizes @xmath20 , by and it is enough to show that @xmath1425 . but this element is of extremal weight and is easily seen to be in the @xmath1115-component of @xmath1050 by direct computation .
[ pp : auto col ] @xmath1290 is stable under @xmath1417 . in particular @xmath1426 as sets .
let @xmath1427 .
write @xmath1428 and @xmath1429 where @xmath1222}$ ] and @xmath1223}$ ] and similarly for @xmath608 . by proposition [ pp : virtual
c ] @xmath1300 is a tableau and @xmath1430 .
suppose first that @xmath898 . by direct computation @xmath1431 where @xmath1432 means to add or subtract @xmath399 from each entry in the tableau @xmath369 .
applying proposition [ pp : virtual c ] again , it follows that @xmath1433 since @xmath1434 has the same above properties that @xmath1300 does .
otherwise let @xmath1309 .
then @xmath1435 for @xmath266 .
now @xmath1436}={\mathbb{p}}(uv|_{[n+1,2n]})-n={\mathbb{p}}(u_2 v_2)-n= ( u_2-n)(v_2-n)\ ] ] which has two columns of equal length .
this means that @xmath1437}$ ] has two columns of equal length , so that it is equal to its @xmath486 tableau . but
@xmath1438})={\mathbb{p}}(uv|_{[n]})+n= ( u_1+n ) ( v_1+n)$ ] .
therefore the columns of @xmath1439 are given by @xmath1440 and @xmath1441 , and the previous argument goes through .
[ thm : c ] the virtual crystal @xmath25 is the crystal graph @xmath19 of the @xmath1407-module @xmath1408 . by @xcite it is known that there is an irreducible integrable @xmath1407-module @xmath1408 with crystal basis which is isomorphic to @xmath1131 as a @xmath1115-crystal . since the @xmath1421-highest weight vectors in @xmath1131 have distinct weights , it follows by remark [ rem : c auto ] that the map @xmath1417 of proposition [ pp : c brs ] for the crystal basis of the module @xmath1408 is uniquely determined by , , , and where only @xmath1442 are used . by proposition [ pp : auto col ] , the involution @xmath1417 of @xmath1308 restricts to an involution on @xmath25 . as such it satisfies the properties of proposition [ pp : c brs ] , by proposition [ pp : virtual auto ] .
the following result is obtained by combining results in @xcite and @xcite .
[ thm : a2d ] for @xmath1387 , the virtual crystal @xmath25 is isomorphic to the crystal graph @xmath19 of the @xmath1392-module @xmath1408 . in @xcite
a @xmath1392-module @xmath1443 was constructed .
its crystal basis @xmath19 was shown to be isomorphic to @xmath1444 as a @xmath1111-crystal .
this agrees with the decomposition specified by .
combining theorem [ thm : virtual fin ] of @xcite and the explicit action of @xmath242 and @xmath405 computed in @xcite , it follows that @xmath1445 as @xmath1392-crystals .
the structure of @xmath19 of type @xmath1 can be deduced from that of the opposite dynkin labeling @xmath22 . to distinguish various objects defined for @xmath22 we shall use the symbol @xmath1446 .
there is an isomorphism of dynkin diagrams @xmath1447 given by @xmath1409 for @xmath1002 .
this induces an isomorphism of weight lattices @xmath1448 given by @xmath1449 for @xmath1002 .
[ thm : a2 ] for @xmath1323 , the virtual crystal @xmath25 is isomorphic to the crystal graph @xmath19 of the @xmath1350-module @xmath1408 .
the @xmath1392-module @xmath1450 is also the @xmath1350-module @xmath1408
. therefore their crystal bases @xmath1451 and @xmath19 are equal as sets ( but not as weighted crystals ) . by abuse of notation
we write @xmath19 for either crystal , using @xmath1446 to distinguish the two structures on this set .
the isomorphism of dynkin diagrams implies that @xmath1452 on @xmath19 .
therefore there is an automorphism @xmath1453 such that @xmath1454 for all @xmath510 .
we claim that @xmath1455 as @xmath1350-crystals . in light of the above facts and the @xmath1392-crystal isomorphism @xmath1456 ,
it is enough to show that there is a unique bijection @xmath1457 such that holds for all @xmath1458 .
but the map @xmath1459 is unique since @xmath1460 is connected as a virtual @xmath1111-crystal and there is no choice for @xmath1461 by weight considerations .
recall that @xmath1462 .
it is easily verified that the automorphism @xmath1463 of the @xmath936-crystal @xmath1308 satisfies .
in this section we prove the fermionic formulas associated with crystals of type @xmath21 , @xmath1 and @xmath2 of the form @xmath1464 as conjectured in @xcite and conjecture a new fermionic formula of type @xmath22 .
we use the virtual crystals discussed in section [ sec : three types ] and results @xcite on bijections between paths of type @xmath34 and rigged configurations .
let @xmath20 be a virtual crystal of type @xmath21 , @xmath1 or @xmath2 as defined in section [ subsec : vc ] .
denote by @xmath1465 the inclusion of the virtual crystal into the corresponding ambient crystal .
the image of @xmath1466 is given in propositions [ pp : virtual c ] , [ pp : virtual d ] and [ pp : virtual a2 ] for type @xmath2 , @xmath21 and @xmath1 , respectively .
under the bijection @xmath727 from paths of type @xmath36 to rigged configurations as described in section [ sec : paths rc ] this image can also be explicitly characterized in terms of rigged configurations as stated in the theorem below .
this theorem will be essential in the proof of the fermionic formulas .
let @xmath588 be a sequence of rectangles where @xmath781 has @xmath782 rows and @xmath783 columns .
define @xmath1467 , @xmath1468 , and let @xmath1469 denote the inclusion @xmath1470 .
let @xmath1471 be the sequence of rectangles corresponding to the underlying tensor product of crystals @xmath1472 ( see ) .
[ thm rc prop ] let @xmath588 be a sequence of single columns where @xmath781 has height @xmath782 and @xmath1473 the corresponding virtual crystal of type @xmath40 , @xmath931 or @xmath923 .
the image @xmath1474 of @xmath1475 is characterized by the set of rigged configurations @xmath703 satisfying the properties : * 1 , 2 , 3 for type @xmath40 * 1 , 3 for type @xmath931 * 1 for type @xmath923 where 1 .
@xmath1476 , i.e. , they are contragredient self - dual .
all parts of @xmath1477 are even .
all riggings in @xmath1478 are even .
the proof of theorem [ thm rc prop ] is given in appendix [ app proof ] .
[ conj : rc prop ] theorem [ thm rc prop ] holds for any sequence of rectangles @xmath37 .
a similar characterization of @xmath1474 seems to exist for type @xmath22 also .
the image of @xmath1466 is described explicitly in proposition [ pp : virtual a2d ] .
[ conj : rc atd ] let @xmath1479 be a virtual crystal of type @xmath22 .
the image @xmath1474 of @xmath1475 is characterized by the set of rigged configurations @xmath703 satisfying the properties : 1 .
@xmath1476 , i.e. , they are contragredient self - dual .
all parts in @xmath1480 are congruent to @xmath67 modulo 2 .
we believe that a proof of this conjecture for tensor products of single columns can be given in a similar fashion to the proof of theorem [ thm rc prop ] as given in appendix [ app proof ] . given a sequence of rectangles @xmath588 and a partition @xmath92 ,
let @xmath1481 be a sequence of partitions with the properties 1 .
@xmath1482 for @xmath1483 2 .
@xmath1477 has only even parts .
define vacancy numbers as follows @xmath1484 denote by @xmath1485 where @xmath1486 the set of @xmath1487 satisfying the two properties listed above and @xmath1488 for all @xmath1483 and @xmath687 .
the set of rigged configurations of type @xmath1486 , denoted by @xmath1489 , is given by @xmath703 such that @xmath1490 and @xmath86 is a double sequence of partitions @xmath1491 such that the partition @xmath1492 lies in a box of width @xmath1493 and height @xmath1494 .
the cocharge for @xmath1495 is defined by @xmath1496 given a sequence of rectangles @xmath588 and a partition @xmath92 , let @xmath1481 be a sequence of partitions such that @xmath1497 define vacancy numbers as follows @xmath1498 denote by @xmath1485 where @xmath1499 the set of @xmath1487 satisfying the constraint above and @xmath1488 for all @xmath1483 and @xmath687 .
the set of rigged configurations of type @xmath1499 , denoted by @xmath1489 , is given by @xmath703 such that @xmath1490 and @xmath86 is a double sequence of partitions @xmath1491 such that the partition @xmath1492 lies in a box of width @xmath1493 and height @xmath1494 .
the cocharge for @xmath1495 is defined by @xmath1500 let @xmath1501 be a composition satisfying @xmath1502 and either @xmath1503 for all @xmath67 or @xmath1504 for all @xmath67 , and @xmath588 a sequence of rectangles .
then let @xmath1481 be a sequence of partitions with the properties @xmath1505 define vacancy numbers as follows @xmath1506 denote by @xmath1485 where @xmath1507 the set of @xmath1487 satisfying the constraint above and @xmath1488 for all @xmath1483 and @xmath687 .
the set of rigged configurations of type @xmath1507 , denoted by @xmath1489 , is given by @xmath703 such that @xmath1490 and @xmath86 is a double sequence of partitions @xmath1491 such that the partition @xmath1492 lies in a box of width @xmath1493 and height @xmath1494 .
the cocharge for @xmath1495 is defined by @xmath1508 it follows from theorem [ thm statistics ] and proposition [ charge energy ] that @xmath1509 by theorem [ thm rc prop ] ( and conjecture [ conj : rc prop ] ) , , and the definitions of rigged configurations in sections [ sec : rc c ] , [ sec : rc a ] and [ sec : rc d ] we can express the one dimensional sum of type @xmath1510 in terms of rigged configurations @xmath1511 where @xmath92 is the `` partition '' corresponding to the weight @xmath1512 .
that is , if @xmath1513 then @xmath92 is the `` partition '' with columns of height @xmath1514 . in the case of @xmath1507 @xmath1515
yields a column of height @xmath399 and width @xmath1516 ( explaining the quotation marks around partition ) .
the contribution of @xmath1517 to @xmath1518 can be evaluated explicitly by noting that the generating function of partitions in a box of dimension @xmath1519 is the @xmath9-binomial coefficient @xmath1520}{0pt}{}{m+p}{m}}_q = \frac{(q)_{m+p}}{(q)_m(q)_p}\ ] ] for @xmath1521 and zero otherwise , where @xmath1522 .
we also introduce @xmath1523 ( @xmath1483 ) by @xmath1524 then may be recast into the form @xmath1525}{0pt}{}{m_i(\nu^{(a)})+p_i^{(a)}(\nu)}{m_i(\nu^{(a)})}}_{q_a}\ ] ] where @xmath1526 .
this formula is known as the fermionic formula . to compare this with the fermionic formula stated in ( * ? ? ?
* eq.(4.5 ) ) several definitions are necessary .
set @xmath1527 with @xmath975 as in .
let @xmath1528 be the number of rectangles in @xmath37 with @xmath195 rows and @xmath67 columns .
let @xmath1529 ( @xmath1530 ) be the simple roots and fundamental weights of @xmath940 and let @xmath1531 be its weight lattice . to relate @xmath1531 with @xmath128 we introduce a @xmath1532-linear map @xmath1533 by @xmath1534 where @xmath1535 is specified as @xmath1536 note that induces @xmath1537 . with these definitions
we obtain @xmath1538 for all @xmath717 and @xmath1539 , @xmath1540 and @xmath1541 then is equivalent to @xmath1542}{0pt}{}{m_i^{(a)}+p_i^{(a)}}{m_i^{(a)}}}_{q_a}\ ] ] where the sum is over all @xmath1543 subject to the constraint and @xmath1544 is given by .
this is ( * ? ? ?
* eq . ( 4.5 ) ) with @xmath1545 and @xmath1546 .
there are also level - restricted versions of the one dimensional configuration sums which are @xmath9-analogues of fusion coefficients . in (
* eq . ( 4.8 ) ) and ( * ? ? ?
3.1 and eq .
( 4.5 ) ) fermionic formulas for the level - restricted configuration sums were conjectured for @xmath1547 . for type @xmath40 these and generalizations
thereof for general @xmath92 are proven in @xcite using the virtual crystal techniques developed in this paper .
similar to the derivation of from theorem [ thm rc prop ] and conjecture [ conj : rc prop ] , a fermionic formula of type @xmath22 can be derived from conjecture [ conj : rc atd ] which to our knowledge is new .
namely , we conjecture that for @xmath428 a crystal of type @xmath22 and @xmath1512 a dominant integral weight @xmath1548}{0pt}{}{m^{(a)}_i+p^{(a)}_i}{m^{(a)}_i}}\\ \prod_{\substack{i\ge 1\\ \text{$i$ even } } } { \genfrac{[}{]}{0pt}{}{m^{(n)}_i+p^{(n)}_i}{m^{(n)}_i } } \prod_{\substack{i\ge 1\\ \text{$i$ odd } } } q^{m^{(n)}_i/2 } { \genfrac{[}{]}{0pt}{}{m^{(n)}_i+p^{(n)}_i-1}{m^{(n)}_i}}.\end{gathered}\ ] ] here @xmath1549 is as in and @xmath1544 as in with @xmath1550 .
the sum is over all @xmath1543 subject to constraint with @xmath1551 and @xmath1552 for @xmath1553 .
note that a summand in is only nonzero if @xmath1554 for all @xmath1483 , @xmath717 and @xmath1555 if @xmath1556 and @xmath67 is odd .
first we show that if @xmath1557 then @xmath703 has the properties of the theorem .
later we will show that the converse holds , that is , if @xmath703 satisfies the the properties then @xmath1557 . by propositions [ pp : virtual c ] , [ pp : virtual d ] and [ pp : virtual a2 ] the image of @xmath1558 yields paths that are contragredient self - dual up to application of the combinatorial @xmath37-matrix for all three types @xmath21 , @xmath1 and @xmath2 .
it was shown in @xcite that under the bijection @xmath727 the combinatorial @xmath37-matrix on lr tableaux yields the identity map on rigged configurations .
hence point 1 follows from theorem [ cdual rc ] in all three cases . to prove point 3 for type @xmath1 and points 2 and 3 for type @xmath2 we proceed by induction on @xmath220 .
let @xmath1559 .
by induction @xmath1560 satisfies point 3 for @xmath1 and points 2 and 3 for @xmath2 where @xmath1561 .
let @xmath1562 be the height of @xmath815 . for type @xmath2 , @xmath1563 for some one column tableaux @xmath362 and @xmath608 of height @xmath1329 and @xmath411 respectively where @xmath1564 . for type @xmath1 , @xmath1565 for some @xmath1566 by theorem [ thm : a2 decomp ] .
let @xmath855 for @xmath1567 ( resp .
@xmath870 for @xmath1568 ) be the lengths of the selected strings corresponding to the @xmath67-th letter in @xmath608 ( resp .
@xmath362 ) under @xmath727 as described in section [ sec : paths rc ] . by definition and by lemma [ lem ineq ]
the following inequalities hold @xmath1569 since @xmath1570 , the number of boxes added to @xmath1477 by @xmath608 is @xmath1571}|$ ] . similarly , the boxes added to the central partition by @xmath362 is @xmath1572}|$ ] which is equal to @xmath529 by proposition [ pp : virtual c ] .
let @xmath195 ( resp .
@xmath54 ) be such that @xmath1573 ( resp .
@xmath1574 ) add boxes to @xmath1477 . since by and @xmath1575 and @xmath1576 it suffices to show that @xmath1577 hence in this case point 2 follows directly by induction for type @xmath2 .
since the vacancy numbers @xmath1578 are all even by the symmetry of point 1 and the fact that under the embedding all rectangles in @xmath1471 of height @xmath399 have even widths , point 3 is also satisfied . for type @xmath1 equation also ensures that after the addition of @xmath1340 all riggings are even since the vacancy numbers are even . the step @xmath1579 is obtained from @xmath1340 by the application of @xmath1580 . by theorem [ thm : emb ]
this induces @xmath1581 on rigged configurations which adds singular strings of length 1 . again , since the @xmath1582 do not change the vacancy numbers and all vacancy numbers are even , point 3 is satisfied for type @xmath1 .
it remains to prove .
to prove we need to study the properties of @xmath855 and @xmath870 more closely . to this end
it is useful to think of @xmath855 and @xmath870 in geometrical terms .
draw dots , called bolts , at @xmath1583 in the @xmath1584-plane whenever @xmath1585 is defined and not infinity .
connect two bolts @xmath1583 and @xmath1586 by a horizontal line from @xmath1587 to @xmath1583 and a vertical line from @xmath1586 to @xmath1587 .
for fixed @xmath67 call this the @xmath67-th @xmath153-slide . in the same fashion
define the @xmath67-th @xmath395-slide by replacing @xmath855 by @xmath870 everywhere .
an example is given in figure [ fig paths ] .
( 180,120)(-20,-20 ) ( 0,0)(0,100 ) ( -2,0)(2,0 ) ( -2,20)(2,20 ) ( -2,40)(2,40 ) ( -2,60)(2,60 ) ( -2,80)(2,80 ) ( -10,0)[]0 ( -10,20)[]1 ( -10,40)[]2 ( -10,60)[]3 ( -10,80)[]4 ( 0,0)(160,0 ) ( 0,-2)(0,2 ) ( 20,-2)(20,2 ) ( 40,-2)(40,2 ) ( 60,-2)(60,2 ) ( 80,-2)(80,2 ) ( 100,-2)(100,2 ) ( 120,-2)(120,2 ) ( 140,-2)(140,2 ) ( 20,-10)[]-3 ( 40,-10)[]-2 ( 60,-10)[]-1 ( 80,-10)[]0 ( 100,-10)[]1 ( 120,-10)[]2 ( 140,-10)[]3 ( 60,0)2blackblack ( 80,20)2blackblack ( 100,40)2blackblack ( 120,40)2blackblack ( 140,80)2blackblack ( 60,0)(60,20)(60,20)(80,20 ) ( 80,20)(80,40)(80,40)(100,40 ) ( 100,40)(120,40 ) ( 120,40)(120,80)(120,80)(140,80 ) call the set of all points @xmath1588 of a slide @xmath1589 with @xmath1590 ( resp .
@xmath1591 ) the positive ( resp .
negative ) part of @xmath1589 .
let us now define a folding operation on slides by mapping the point @xmath1588 of a slide @xmath1589 to @xmath1592 .
a bolt that changes under folding transforms to a `` nut '' .
let us now define nut - bolt pairs as follows . by the symmetry of the rigged configuration @xmath703 and the property of @xmath608 that there is no @xmath1593 with @xmath1594 ( otherwise the duality property of proposition [ pp : virtual c ] can not hold ) the @xmath54-th folded @xmath153-slide must cross each @xmath180-th folded @xmath153-slide with @xmath1595 and @xmath1596 in at least one bolt . if @xmath1597 with @xmath1598 then call @xmath1599 $ ] a nut - bolt pair ( note that here we also allow @xmath1600 ) . by induction on @xmath687
the @xmath1601-th folded @xmath153-slide must cross each @xmath180-th folded @xmath153-slide with @xmath1602 and @xmath1603 in at least one unclaimed bolt .
call @xmath1604 $ ] a nut - bolt pair if @xmath1605 and @xmath1606 $ ] is not a nut - bolt pair .
an example is given in figure [ fig nut - bolt ] .
[ lem nb ] for @xmath687 and @xmath1598 the following must be nut - bolt pairs : @xmath1607
\quad \text{for some $ k$ } \qquad \text{or } \qquad [ s_{a+i - j}^{(n - j)},\ell_{b+i}^{(n+j)}];\\ \label{nb 2 } & [ s_k^{(n - j)},s_{a+i}^{(n+j ) } ] \quad \text{for some $ k$ } \qquad \text{or } \qquad [ \ell_{b+i - j}^{(n - j)},s_{a+i}^{(n+j)}].\end{aligned}\ ] ] if @xmath1608 $ ] is not a nut - bolt pair then @xmath1609 and similarly if @xmath1610 $ ] is not a nut - bolt pair then @xmath1611 the proof of these lemmas is given by an intertwined induction on @xmath67 and @xmath191 . we will give the proofs separately , but the inductive step of each will rely on the validity of both other lemmas for either the same @xmath67 and greater @xmath191 , or smaller @xmath67 .
we prove by induction on increasing @xmath67 and decreasing @xmath191 .
suppose that for given @xmath67 and @xmath191 @xmath1604 $ ] is not a nut - bolt pair for any @xmath180 .
we want to show that then @xmath1608 $ ] is a nut - bolt pair .
first assume that @xmath191 is maximal , that is @xmath1615 .
note that from the duality and the tableau condition of @xmath1340 as stated in proposition [ pp : virtual c ] it follows that @xmath1616 and @xmath1617 for @xmath1615 . since we know already that the final rigged configuration has to be symmetric , i.e. satisfies condition 1 of theorem [ thm rc prop ] , and since @xmath1618 for all @xmath1619 by we therefore must have @xmath1620 so that @xmath1608 $ ] forms a nut - bolt pair .
assume that @xmath1623 $ ] is a nut - bolt pair .
if @xmath1624 $ ] is also a nut - bolt pair then by symmetry @xmath1620 so that @xmath1608 $ ] indeed forms a nut - bolt pair . if @xmath1613 $ ] is a nut - bolt pair for some @xmath180 or @xmath1625 then @xmath1626 since otherwise @xmath1604 $ ] would be a nut - bolt pair which contradicts our assumptions . then again by symmetry
@xmath1627 and @xmath1608 $ ] forms a nut - bolt pair
. otherwise assume that @xmath1628 $ ] is a nut - bolt pair for some @xmath180 .
if @xmath1606 $ ] is also a nut - bolt pair then @xmath1629 $ ] can not be a nut - bolt pair for any @xmath1630 since otherwise @xmath1631 $ ] would be a nut - bolt pair which contradicts our assumptions . hence @xmath1624 $ ]
must be a nut - bolt pair . by the change of vacancy numbers
there are no singular strings of length @xmath1632 in the @xmath1633-th rigged partition after the addition of all letters up to @xmath1634 . since @xmath1635 this shows by symmetry that @xmath1636 and hence @xmath1608 $ ] forms a nut - bolt pair .
if @xmath1606 $ ] is not a nut - bolt pair then @xmath1637 by lemma [ lem ess ] ( note that @xmath1638 since otherwise @xmath1604 $ ] would be a nut - bolt pair which contradicts our assumptions ) . by the same arguments as before there are no singular strings of length @xmath1632 in the @xmath1633-th rigged partition after the addition of all letters up to @xmath1634 .
using @xmath1639 follows again .
fix @xmath67 .
let @xmath191 be maximal , that is @xmath1615 .
assume that @xmath1629 $ ] is a nut - bolt pair for some @xmath1640 .
the conditions of the lemma require that @xmath1641 . by definition @xmath1642
is the length of the largest singular string in @xmath1643 subject to the condition @xmath1644 .
there is a singular string of length @xmath1645 in @xmath1646 .
hence by symmetry @xmath1631 $ ] must form a nut - bolt pair so that @xmath1647 .
this contradicts @xmath1641 .
therefore , by , either @xmath1625 or @xmath1624 $ ] is a nut - bolt pair . by the definition of @xmath1642
there is no singular string of length @xmath1648 in @xmath1643 and hence by symmetry in @xmath1646 . using that @xmath1649 and taking into account the change in vacancy numbers this implies that @xmath1650 . if @xmath1621 then by and and the crossing conditions on @xmath153-slides exactly one of the following three must be a nut - bolt pair for @xmath1651 : ( 1 ) @xmath1623 $ ] ; ( 2 ) @xmath1628 $ ] ; ( 3 ) @xmath1652 $ ] .
we are going to treat each case separately .
assume that case ( 1 ) holds .
if @xmath1653 , then by the definition of @xmath1642 there are no singular strings of length @xmath1654 in @xmath1643 . by symmetry
there are no singular strings of length @xmath1655 in @xmath1646 .
taking into account the change of vacancy number , there are still no singular strings of length @xmath1656 in the @xmath1633-th rigged partition after the addition of all letters up to @xmath1634 .
since @xmath1657 by it follows that @xmath1650 .
if @xmath1658 , then @xmath1624 $ ] must be a nut - bolt pair . by the same arguments as above
there are no singular strings of length @xmath1659 in the @xmath1633-th rigged partition after the addition of all letters @xmath1634 .
since @xmath1660 by this implies @xmath1650 .
assume that case ( 2 ) holds .
then the conditions of lemma [ lem ess ] are satisfied for @xmath191 replaced by @xmath1622 and by induction @xmath1661 .
note that @xmath1662 in case ( 2 ) . since @xmath1663 it follows that @xmath1664 .
hence by it follows that @xmath1665 .
assume that case ( 3 ) holds .
let us first show that @xmath1666 . if @xmath1667 $ ] is not a nut - bolt pair then by induction @xmath1668 ( note that @xmath1669 since otherwise @xmath1670 $ ] would be a nut - bolt pair which contradicts our assumptions ) .
since @xmath1671 by it follows that @xmath1672 .
if @xmath1667 $ ] is a nut - bolt pair then ( a ) @xmath1624 $ ] or ( b ) @xmath1673 $ ] is a nut - bolt pair . in case ( a ) @xmath1674 so that the assertion holds . in case
( b ) the assumptions of the lemma hold for @xmath67 replaced by @xmath1145 so that by induction @xmath1675 . since @xmath1676 by and the assertion holds .
there are no singular strings of length @xmath1677 in the @xmath1633-th rigged partition after the addition of all letters up to @xmath1634 because of the change of the vacancy numbers .
furthermore there are no singular strings of length @xmath1678 in @xmath1643 by the definition of @xmath1642 .
hence by symmetry and change in vacancy numbers , there are no singular strings of length @xmath1679 in the @xmath1633-th rigged partition after the addition of all letters up to @xmath1634 .
now assume that @xmath1672 . then these conditions and the fact that @xmath1624 $ ] is a nut - bolt pair if @xmath1680 imply that @xmath1650 .
hence the only case left to consider is @xmath1681 . by the above arguments this case can only occur when @xmath1667 $ ] and @xmath1673 $ ] are nut - bolt pairs with @xmath1675 .
we will show that @xmath1681 can not occur unless @xmath1682 in which case lemma [ lem ess ] obviously holds .
assume that @xmath1683 . by the definition of @xmath1684 and symmetry
there is no singular string of length @xmath1685 available for @xmath1686 unless ( i ) @xmath1687 or ( ii ) @xmath1688 .
assume case ( i ) holds . by either @xmath1689
$ ] or @xmath1690 $ ] has to be a nut - bolt pair . however , the first case yields a contradiction since @xmath1691 so that @xmath1690 $ ] must be a nut - bolt pair .
assume that case ( ii ) holds .
then @xmath1692 $ ] must be a nut - bolt pair . by this
implies that @xmath1693 and @xmath1690 $ ] must be a nut - bolt pair .
hence in both case ( i ) and case ( ii ) @xmath1690 $ ] must be a nut - bolt pair and @xmath1694 .
now , there is no singular string of length @xmath1685 for @xmath1686 unless @xmath1695 .
but then @xmath1696 $ ] must be a nut - bolt pair since otherwise @xmath1697 $ ] is a nut - bolt pair which contradicts our assumptions .
this in turn implies that @xmath1698 . by either @xmath1699
$ ] or @xmath1692 $ ] must be a nut - bolt pair .
however , the first case contradicts @xmath1681 since @xmath1700 . hence @xmath1692 $ ]
must be a nut - bolt pair . by induction @xmath1701 . in summary
both @xmath1696 $ ] and @xmath1692 $ ] are nut - bolt pairs with @xmath1702 .
now repeat the entire argument which shows that both @xmath1703 $ ] and @xmath1704 $ ] are nut - bolt pairs with @xmath1705 etc .
since there are only finitely many @xmath153-slides these conditions must eventually break down which shows that @xmath1681 can not occur when @xmath1683 .
this concludes the proof of lemma [ lem ess ] . if @xmath1706 then by proposition [ pp : virtual c ] also @xmath1707 and @xmath1708 .
this implies that @xmath1709 is the largest singular string in @xmath1478 and @xmath1710 is the largest singular string in the central rigged partition after the addition of @xmath608 .
since @xmath1711 is singular it follows that @xmath1712 . after the addition of @xmath608 the vacancy numbers of the strings of length @xmath1713 in the central rigged partition decrease by one
however , since all labels are even no string of length @xmath1714 becomes singular .
hence @xmath1715 .
if @xmath1717 $ ] is a nut - bolt pair then by @xmath1718 .
furthermore @xmath1719 since otherwise the folding of the @xmath54-th @xmath153-slide would not cross the @xmath1720-th @xmath153-slide in a bolt .
this implies that @xmath1721 .
also @xmath1722 since otherwise @xmath1717 $ ] was not a nut - bolt pair .
since there is a singular string of length @xmath1711 this implies that @xmath1712 . after the addition of @xmath608 ,
the vacancy numbers corresponding to the strings of length @xmath1723 in the central rigged partition have decreased by 1 .
again , since there are no singular strings of length @xmath1723 in @xmath1478 by the definition of @xmath1709 and all labels are even , it follows that there are no singular strings of length @xmath1724 after the addition of @xmath608 .
hence @xmath1715 .
if @xmath1725 $ ] is a nut - bolt pair then by and we have @xmath1726 and @xmath1721 .
this implies @xmath1712 . after the addition of @xmath608
the vacancy number in the central rigged partition corresponding to the strings of length @xmath1723 decreases by one and of length @xmath1727 increases by one .
hence there are no singular strings of length @xmath1728 and by the now familiar arguments also not for @xmath1724 since all labels are even .
this proves @xmath1715 .
finally let @xmath1729 $ ] be a nut - bolt pair .
note that this case can only happen if @xmath1730 .
we will show that @xmath1731 .
since by @xmath1732 it then follows that @xmath1715 . if @xmath1733 then @xmath1734 for @xmath1729 $ ] to hold .
the length of the longest singular string in @xmath1735 is @xmath1709 and hence by symmetry also in @xmath1736 .
this forces @xmath1731 .
hence assume @xmath1737 .
if @xmath1738 $ ] is a nut - bolt pair then @xmath1731 holds by lemma [ lem ess ] .
if @xmath1739 $ ] is a nut - bolt pair then the vacancy numbers in the @xmath1740-st rigged partition corresponding to strings of length @xmath1741 are increased by one after the addition of @xmath608 so that there are no singular strings of this length . by lemma [ lem ess ] @xmath1742 (
note that the case @xmath1743 is included here ) . hence using @xmath1744 and
this implies @xmath1745 . by the definition of @xmath1746
there are no singular strings in @xmath1735 of length @xmath1747 and hence by symmetry also not in @xmath1736 . after the addition of @xmath608
there is a singular string of length @xmath1748 so that @xmath1731 . to conclude assume that @xmath1749 $ ] is a nut - bolt pair . then by and @xmath1750 . by the same arguments as in the previous case
there are no singular strings of length @xmath1751 in the @xmath1740-st rigged partition after the addition of @xmath608 so that @xmath1731 as asserted .
assume that @xmath1753 $ ] is a nut - bolt pair .
then @xmath1754 , @xmath1755 and @xmath1756 . since there is a singular string of length @xmath1757 after the addition of @xmath608 and the letters @xmath1758 it follows that @xmath1759 by . by the definition of @xmath1757
there are no singular strings of length @xmath1760 in @xmath1478 .
the vacancy numbers of the central rigged partition corresponding to the strings of length @xmath1760 decrease by one . hence due to the even labels there are no singular strings of length @xmath1761 in the central rigged partition after the addition of all letters up to @xmath1634 .
this proves .
now assume that @xmath1762 $ ] is a nut - bolt pair . by induction @xmath1763 .
the vacancy numbers of the central rigged partition after the addition of all letter up to @xmath1634 for the strings of length @xmath1764 increase by one and for the strings of length @xmath1765 decrease by one . since there are no singular strings of length @xmath1766 in @xmath1478 by the definition of @xmath1757 there are no singular strings of length @xmath1767 in the central rigged partition after the addition of all letters up to @xmath1634 due to the even labels . also @xmath1768 . hence holds unless @xmath1769 and @xmath1770 .
we will show that in the present case this can not occur .
assume that @xmath1769 and @xmath1770 .
this requires @xmath1771 and @xmath1772 by . by either ( 1 ) @xmath1773 $ ] or ( 2 ) @xmath1774 $ ] forms a nut - bolt pair .
since @xmath1762 $ ] is a nut - bolt pair by assumption , @xmath180 in case ( 2 ) must be @xmath1775 .
hence if case ( 2 ) holds there is no singular string of length @xmath1776 in the central rigged partition after the addition of all letter up to @xmath1634 by the change in vacancy number .
hence @xmath1769 can not hold . by the same arguments there is no singular string of length @xmath1776 if @xmath1777 . hence either ( a ) @xmath1778 or ( b ) @xmath1779 . since @xmath1780 case ( b ) requires case ( 2 ) which we already argued yields a contradiction .
hence assume case ( a ) and case ( 1 ) holds .
but then by and @xmath1781 $ ] must be a nut - bolt pair which contradicts @xmath1782 .
this concludes the proof of when @xmath1762 $ ] is a nut - bolt pair .
finally assume that @xmath1783 $ ] is a nut - bolt pair .
if @xmath1784 then @xmath865 and @xmath1785 , @xmath1786 and @xmath1787 . by symmetry
@xmath1788 $ ] forms a nut - bolt pair which contradicts our assumption .
hence we may assume from now on that @xmath1789 .
we will show that @xmath1790 . then by and it follows that @xmath1791 . if @xmath1792 $ ] is a nut - bolt pair then @xmath1793 since otherwise the @xmath1794-st and the @xmath1601-th @xmath153-slide would not cross in a bolt .
hence by the definition of @xmath1795 there are no singular strings of length @xmath1796 in @xmath1735 . by symmetry
there are no singular strings of length @xmath1796 in @xmath1736 .
since @xmath1797 by it follows that @xmath1790 . if @xmath1798 $ ] is a nut - bolt pair then @xmath1790 follows from lemma [ lem ess ] .
the last case to consider is that @xmath1799 $ ] is a nut - bolt pair .
note that due to the change of vacancy numbers there are no singular strings of length @xmath1800 in the @xmath1740-st rigged partition after the addition of all letters up to @xmath1634 . also by the definition of @xmath1795 and symmetry there are no singular strings of length @xmath1801 in the same rigged partition .
hence , if @xmath1802 then @xmath1803 as asserted . we may therefore restrict our attention to the case @xmath1804 with @xmath1805 ( note that @xmath1806 would imply the assertion ) .
we will show that these conditions can not hold simultaneously . by lemma [ lem ess ]
@xmath1804 can only happen if ( 1 ) @xmath1807 $ ] is a nut - bolt pair or ( 2 ) @xmath1808 .
note that case ( 2 ) implies case ( 1 ) since otherwise @xmath1762 $ ] is a nut - bolt pair which contradicts our assumptions .
case ( 1 ) implies by induction that @xmath1809 . in summary
@xmath1807 $ ] is a nut - bolt pair with @xmath1810 .
there are no singular strings of length @xmath1811 available in the @xmath1740-st rigged partition for @xmath1812 unless ( a ) @xmath1813 or ( b ) @xmath1814 . in case ( a ) by either
@xmath1815 $ ] or @xmath1816 $ ] forms a nut - bolt pair .
the first case contradicts @xmath1817 . in case ( b ) @xmath1818 so that @xmath1816 $ ] must form a nut - bolt pair .
hence in both case ( a ) and ( b ) @xmath1816 $ ] forms a nut - bolt pair and @xmath1818 . by the change in vacancy numbers there is no singular string of length @xmath1811 available for @xmath1812 unless @xmath1819 . then @xmath1820 $ ] must be a nut - bolt pair since otherwise @xmath1821 $ ] forms a nut - bolt pair which contradicts our assumptions . by induction @xmath1822 . repeating the same arguments
shows that @xmath1823 $ ] forms a nut - bolt pair with @xmath1824 etc .
since there are only finitely many @xmath153-slides these conditions must be eventually broken which shows that @xmath1804 is not possible unless @xmath1825 .
this concludes the proof of .
so far we have shown that the conditions of theorem [ thm rc prop ] are satisfied if @xmath1557 .
it remains to show the converse , that is , if @xmath703 satisfies the conditions of theorem [ thm rc prop ] then @xmath1557 .
let @xmath1826 where @xmath1827 if @xmath781 is a single column of height @xmath1828 and @xmath1829 is of type @xmath2 , @xmath1830 if @xmath781 is a single column of height @xmath1828 and @xmath1829 of type @xmath21 , or @xmath1831 otherwise . by theorem [ cdual rc ] and
the fact that the combinatorial @xmath37-matrix on paths yields the identity map on rigged configurations , condition 1 of theorem [ thm rc prop ] ensures that @xmath638 is contragredient self - dual , that is @xmath1832 . for type @xmath21
this already shows that @xmath1557 by proposition [ pp : virtual d ] . for type @xmath1 act on @xmath703 with @xmath1833 with @xmath411 as small as possible .
that is , first remove singular strings of length 1 from the @xmath180-th rigged partition for @xmath1834 if possible .
repeat this for @xmath1835 until no more such singular strings can be removed for @xmath1836 . by induction
we may assume that @xmath1837 is in @xmath1838 if it satisfies the properties of theorem [ thm rc prop ] where @xmath1561 .
set @xmath1839 and @xmath1840 for type @xmath2 and @xmath1841 for type @xmath1 .
then by proposition [ pp : virtual c ] and theorem [ thm : a2 decomp ] it suffices to prove that @xmath1842}$ ] is a tableau , @xmath1843}|-|v|_{[n]}|=n - r$ ] and @xmath1844 satisfies point 3 of theorem [ thm rc prop ] for type @xmath1 and points 2 and 3 for type @xmath2 . the rigged configuration @xmath1844 is obtained from @xmath703 by the inverse algorithm to the one described in section [ sec : paths rc ] .
let @xmath1845 for @xmath1846 and @xmath1847 for @xmath1848 be rigged configurations where @xmath1849 , @xmath1850 and @xmath1851 .
the rigged configuration @xmath1852 is obtained from @xmath1845 in the following way : select singular strings of length @xmath1853 in the @xmath67-th , @xmath1854-th , rigged partition in @xmath1845 minimal such that @xmath1855 . if no such singular string exists in the @xmath180-th rigged partition set @xmath1856 .
let @xmath1857 be minimal such that @xmath1858 . then remove a box from each of the selected strings in the @xmath67-th to @xmath1859-th rigged partition in @xmath1845 keeping them singular and leaving all other strings unchanged .
the result is @xmath1852 .
define @xmath1860 , @xmath1861 and @xmath1862 in the analogous fashion starting with @xmath1847 .
then let @xmath362 be the column tableau of height @xmath1329 with entries @xmath1863 and @xmath608 the column tableau of height @xmath411 with entries @xmath1864 .
similarly to the following holds .
let @xmath195 and @xmath1865 be defined such that @xmath1866 are precisely all finite @xmath1867 .
define @xmath54 and @xmath1868 similarly in terms of @xmath1869 .
then similar to the proof of it can be shown that @xmath1870 and @xmath1871 ( for type @xmath2 , @xmath1872 since by property 2 all parts in @xmath1477 are even . for type
@xmath1 observe that @xmath1872 also due to the application of @xmath1873 to @xmath703 ) .
this proves in particular that @xmath1844 satisfies point 3 for type @xmath1 and points 2 and 3 of theorem [ thm rc prop ] for type @xmath2 .
it also shows that @xmath1843}|-|v|_{[n]}|=(n - d)-(r - d)=n - r$ ] as desired . finally , lemma [ lem nb ] holds with @xmath870 and @xmath855 replaced by @xmath1874 and @xmath1860 , respectively .
the proof is similar to the proof of lemma [ lem nb ] . from this
follows that @xmath1875 . in particular
, this implies that @xmath1876 if @xmath1877 .
let @xmath120 be minimal such that @xmath1878 , that is , @xmath1879 .
then the above shows that @xmath1880 by the contragredient self - duality @xmath1842}$ ] is a tableau if and only if @xmath1881}$ ] is a tableau .
now assume that @xmath1881}$ ] is not a tableau , that is , there exists an index @xmath67 such that @xmath1882 for all @xmath1883 and @xmath1884 .
let @xmath1885 .
then it follows from duality that @xmath1886 .
this contradicts .
hence @xmath1842}$ ] must be a tableau as asserted .
a. n. kirillov and n. y. reshetikhin , _ representations of yangians and multiplicities of the inclusion of the irreducible components of the tensor product of representations of simple lie algebras _ , j. soviet math .
* 52 * ( 1990 ) , no .
3 , 31563164 .
i. g. macdonald , _ symmetric functions and hall polynomials _ , second edition , with contributions by a. zelevinsky , oxford mathematical monographs , oxford science publications , the clarendon press , oxford university press , new york , 1995 . | we introduce `` virtual '' crystals of the affine types @xmath0 , @xmath1 and @xmath2 by naturally extending embeddings of crystals of types @xmath3 and @xmath4 into crystals of type @xmath5 .
conjecturally , these virtual crystals are the crystal bases of finite dimensional @xmath6-modules associated with multiples of fundamental weights .
we provide evidence and in some cases proofs of this conjecture .
recently , fermionic formulas for the one dimensional configuration sums associated with tensor products of the finite dimensional @xmath6-modules were conjectured by hatayama et al .
we provide proofs of these conjectures in specific cases by exploiting duality properties of crystals and rigged configuration techniques . for type @xmath1
we also conjecture a new fermionic formula coming from a different labeling of the dynkin diagram . partially supported by nsf grant dms-9800941 . |
we thank s .-
zhu for helpful discussions .
this work was supported by the national natural science foundation of china under grants no .
11274061 and no .
j.m.h conceived and supervised the project . j.m.h . and w.c .
made the calculations and wrote the paper .
* supplementary information * accompanies this paper . + + * competing financial interests : * the authors declare no competing financial interests .
+ * hidden symmetry and protection of dirac points on the honeycomb lattice supplementary information * jing - min hou@xmath86 and wei chen@xmath87 _ @xmath86department of physics , southeast university , nanjing 211189 , china _
_ @xmath87college of science , nanjing university of aeronautics and astronautics , nanjing 210016 , china _
for the general honeycomb lattice with the bond angle @xmath0 , the tight - binding hamiltonian can be written as , @xmath88 + h.c .
\label{thh}\end{aligned}\ ] ] where @xmath89 is the annihilation operator that destructs a particle in the wannier state located at the site @xmath90 in sublattice @xmath91 and @xmath92 is the annihilation operator that destructs a particle in the wannier state located at the site @xmath93 in sublattice @xmath94 ; the subscript @xmath95 is the coordinate for the lattice sites ; @xmath96 represent the unit vectors between the two nearest lattice sites ; @xmath4 and @xmath5 are the amplitudes of hopping as shown in fig.1 a in the main text .
we take the fourier s transformation to the annihilation operators as @xmath97 where @xmath98 and @xmath99 represent the positions of lattice sites in sublattice lattice @xmath91 and @xmath94 , respectively . substituting eqs.([fta ] ) and ( [ ftb ] ) into equation ( [ thh ] ) , we obtain @xmath100+h.c.\end{aligned}\ ] ] we define the two - component annihilation operator as @xmath101^t$ ] and the hamiltonian can be written as @xmath102 . here
, @xmath103 is the bloch hamiltonian of the honeycomb lattice model for the wave vector @xmath37 as @xmath104\sigma_x+[2t_1\cos(\cos\theta k_x)\sin(\sin\theta k_y)-t_2\sin k_y]\sigma_y\label{bhh2}\end{aligned}\ ] ] where @xmath6 and @xmath7 are the pauli matrices .
this bloch hamiltonian is equation ( 1 ) in the main text .
the corresponding dispersion relation is @xmath105+t_2 ^ 2}\end{aligned}\ ] ] for the ideal honeycomb lattice , @xmath1 , such as graphene , the bloch hamiltonian is @xmath106\sigma_x+[2t_1\cos ( { \sqrt{3 } } k_x/2)\sin ( k_y/2)-t_2\sin k_y]\sigma_y\end{aligned}\ ] ] and the dispersion relation is @xmath107 the honeycomb lattice with @xmath2 is the brick - wall lattice . substituting @xmath2 into equation ( [ bhh2 ] ) , we obtain the bloch hamiltonian for the brick - wall lattice as @xmath108\sigma_x -t_2\sin k_y\sigma_y\label{shb}\end{aligned}\ ] ] the corresponding dispersion relations is @xmath109 for the square lattice with a hopping - accompanying @xmath8 phase , the tight - binding hamiltonian can be written as @xmath110 + h.c.\end{aligned}\ ] ] where @xmath24 and @xmath111 represent the unit vectors in the @xmath112 and @xmath113 directions , respectively ; @xmath10 and @xmath11 are the amplitudes of hopping along the @xmath112 and @xmath113 directions , respectively . taking the fourier s transformation , we obtain the hamiltonian as @xmath114+h.c.\end{aligned}\ ] ] we rewritten the hamiltonian as @xmath115 with @xmath101^t$ ] .
then , the bloch hamiltonian for the square lattice model is @xmath116 the corresponding dispersion relation is @xmath117
for all the lattices , we assume that the bond length @xmath118 . for the general honeycomb lattice with the bond angle @xmath0 , the primitive lattice vectors are @xmath119 and @xmath120 .
the primitive reciprocal lattice vectors are @xmath121 and @xmath122 . for @xmath1 case , such as graphene ,
the symmetric brillouin zone is hexagon , i.e. , the area enclosed by the black lines in fig.[figs1 ] a. an alternative brillouin zone equivalent to the symmetric brillouin zone is a diamond , i.e. , the yellow shaded area in fig.[figs1 ] a. in our work , for convenience , we always use the diamond brillouin zone for the honeycomb lattice
. for and the square lattice , the primitive lattice vectors are @xmath123 and @xmath124 .
the primitive reciprocal lattice vectors are @xmath125 and @xmath126 .
the square lattice has a square brillouin zone as shown in fig.[figs1 ] b. the brick - wall lattice can be considered a special honeycomb lattice with the bond angle @xmath2 .
the primitive lattice vectors become @xmath123 and @xmath124 .
the primitive reciprocal lattice vectors are @xmath125 and @xmath126 .
the corresponding brillouin zone turns into a square , which is the same with that of the square lattice as shown in fig.[figs1 ] b.
the square lattice with a hopping - accompanying @xmath8 phase respects a hidden symmetry , which is defined as @xmath127 where @xmath23 , @xmath25 , @xmath6 , and @xmath26 represent a translation along the @xmath112 direction by a unit vector , the complex conjugation , the sublattice exchange , and a local @xmath27 gauge transformation , respectively . from fig.[symmetry ] , we can find the square lattice with a hopping - accompanying @xmath8 phase is invariant under the actions of @xmath128 , @xmath25 , @xmath6 , and @xmath26 in order .
thus , we conclude that the square lattice with a hopping - accompanying @xmath8 phase respects the hidden symmetry @xmath28 .
we can interpret the mapping @xmath48 in an intuitive way . to this end , we divide it into two mappings in order as @xmath129 . here , @xmath130 is the mapping from the general honeycomb lattice model with the bond angle @xmath0 into the brick - wall lattice model and @xmath131 is the mapping from the brick - wall lattice model with the hopping amplitude ratio @xmath14 into the square lattice model . in the following ,
we explain the two mappings in details .
the general honeycomb lattice model with the bond angle @xmath0 is equivalent with the brick - wall lattice model in some sense . to manifest this equivalence
, we define a mapping @xmath130 from the general honeycomb lattice model to the brick - wall lattice model as @xmath132 where @xmath47 and @xmath133 are the bloch functions of the honeycomb lattice model and the brick - wall lattice model , respectively . to find the explicit form of the mapping , we take a transformation to the bloch hamiltonian ( equation 1 in the main text ) as @xmath134 , where @xmath135 is the transformation matrix defined as @xmath136 for @xmath137 . here
@xmath138 , @xmath139 , @xmath140 $ ] and @xmath141 $ ] , which satisfy the identity @xmath142 .
when @xmath143 , the transformation matrix @xmath135 is ill - defined .
for the continuity of the mapping , when @xmath143 , we take the limit as the definition of @xmath135 .
we then obtain @xmath144\}\sigma_x -t_2\sin [ ( 1+\sin\theta)k_y]\sigma_y\end{aligned}\ ] ] substituting @xmath145 and @xmath146 into @xmath147 , we obtain @xmath148\sigma_x -t_2\sin p_y\sigma_y \label{bhb2}\end{aligned}\ ] ] which is just the bloch hamiltonian ( equation [ shb ] ) of the brick - wall lattice model .
the mapping @xmath149 has the effect on the bloch functions and the wave vectors as @xmath150 and @xmath151 .
this mapping is one - to - one and surjective .
thus , we can regard this mapping as a kind of equivalence .
the explicit form of the mapping depends on the bond angle @xmath0 .
when @xmath2 , this mapping is an identity mapping .
the mapping @xmath130 gives a one - to - one correspondence between the brillouin zones of the honeycomb lattice and the brick - wall lattice .
that is to say , for some wave vector @xmath37 in the brillouin zone of the honeycomb lattice , the bloch hamiltonian @xmath103 and its bloch functions @xmath47 , there correspondingly exist a wave vector @xmath152 in the brillouin zone of the brick - wall lattice , the bloch hamiltonian @xmath153 , and block functions @xmath133 .
the mapping from the brillouin zone of the honeycomb lattice into that of the brick - wall lattice is schematically shown in fig.[map1 ] .
next , we can define a mapping from the brick - wall lattice model to the square lattice model as @xmath154 where @xmath15 is the hopping amplitude ratio of the brick - wall lattice . to find the explicit form of this mapping ,
we suppose @xmath155 and @xmath156,&\mbox{for}\ k_x\leq0\\ \arccos[\frac{2}{2+\beta}\cos p_x+\frac { \beta } { 2+\beta}\cos p_y ] , & \mbox{for } \ k_x\geq0 \end{cases}\\ k_y&=&p_y\end{aligned}\ ] ] where @xmath152 is a wave vector in the brillouin zone of the brick - wall lattice . through
the mapping @xmath131 , the bloch hamiltonian of the brick - wall lattice ( [ bhb2 ] ) becomes the form as follows , @xmath157 with @xmath50 and @xmath51 , which is just the bloch hamiltonian of the square lattice model .
the explicit form of this mapping depends on the hopping amplitude ratio @xmath14 .
this mapping is not surjective .
that is to say , the image of the mapping for the brillouin zone of the brick - wall lattice just covers part of the brillouin zone of the square lattice .
the mapping for the wave vectors is schematically shown in fig.[map2 ] . in fig.[map2 ] , the left panel shows the brillouin zone of the brick - wall lattice . in order to clearly manifest the mapping from the brick - wall lattice model to the square lattice model , we first map the brillouin zone of the brick - wall lattice into the momentum space of the square lattice , not restricted in the brillouin zone , as shown in the middle panel of fig.[map2 ] .
the image of the brillouin zone of the brick - wall lattice in the momentum space of the square lattice looks like a butterfly .
the left and right halves of the brillouin zone of the brick - wall lattice map into the left and right wings of the butterfly , respectively .
if we restrict the image of the mapping in the brillouin zone of the square lattice , then the butterfly - like image is equivalent to that as shown in the right panel of fig.[map2 ] .
@xmath28 is the operator of the hidden symmetry that is respected by the square lattice considered in the main text .
the @xmath28-invariant points in the brillouin zone are @xmath38 and @xmath39 .
since the square of the hidden symmetry operator @xmath28 is a translation operator as @xmath29 , we have @xmath158 .
we define @xmath159 as the inner product of the two wave functions @xmath160 and @xmath161 .
the anti - unitary operator @xmath28 has the property that @xmath162 .
therefore , at the @xmath28-invariant point @xmath163 in the brillouin zone , we have @xmath164 where @xmath165 is the @xmath112 component of the @xmath28-invariant point @xmath163 and the input of the bloch functions is omitted for convenience . for the @xmath28-invariant points @xmath43 , where @xmath166 , we have @xmath42 , then we obtain the solution @xmath167 , that is to say , @xmath168 and @xmath169 are orthogonal to each other . while , for the @xmath28-invariant points @xmath45 , where @xmath170 , we have @xmath44 , so @xmath171 is unconstrained for equation ( [ prot ] ) . therefore , we arrive at the conclusion that the system must be degenerate at points @xmath43 in the brillouin zone for the square lattice , which is consistent with the dispersion relation calculated .
@xmath58 is a self - mapping of the honeycomb lattice model with the bond angle @xmath0 and the hopping amplitude ratio @xmath14 .
we assume that @xmath47 is the bloch function of the honeycomb lattice model .
since @xmath58 is a self - mapping of the honeycomb lattice model , we have @xmath172 which is also the bloch function of the honeycomb lattice model . after the action of the operator @xmath58 ,
the wave vector @xmath37 becomes @xmath173 where @xmath174 with @xmath54\}$ ] .
if @xmath175 is satisfied , where @xmath176 is a reciprocal lattice vector of the honeycomb lattice , @xmath37 is a @xmath58-invariant point in the brillouin zone of the honeycomb lattice .
we assume @xmath177 is a @xmath58-invariant point in the brillouin zone . substituting equation ( [ trans ] ) and @xmath178 into equation ( [ invariant ] ) , we obtain the following equation @xmath179 solving the above equation , we obtain the @xmath58-invariant points in the brillouin zone are @xmath180 . at the @xmath58-invariant points @xmath181 ,
we have @xmath182 it is easy to verify that @xmath69 .
therefore , we have @xmath70}\psi_{h,\mathbf{k}}(\mathbf{r } ) $ ] . at the @xmath58-invariant point @xmath177
, we have the following equation : @xmath183 } ( { \psi'_{h,\mathbf{q}_i } } , \psi_{{h,\mathbf{q}_i}}).\end{aligned}\ ] ] substituting @xmath59 , we obtain @xmath184}=-1 $ ] at @xmath59 .
thus , we have the solution @xmath185 , which implies that @xmath186 and @xmath187 are orthogonal to each other
. we can conclude that there must be the band degeneracy at the points @xmath59 . | the honeycomb lattice possesses a novel energy band structure , which is characterized by two distinct dirac points in the brillouin zone , dominating most of the physical properties of the honeycomb structure materials .
however , up till now , the origin of the dirac points is unclear yet . here
, we discover a hidden symmetry on the honeycomb lattice and prove that the existence of dirac points is exactly protected by such hidden symmetry .
furthermore , the moving and merging of the dirac points and a quantum phase transition , which have been theoretically predicted and experimentally observed on the honeycomb lattice , can also be perfectly explained by the parameter dependent evolution of the hidden symmetry .
graphene , a honeycomb lattice of carbon atoms , has attracted extraordinary attention in the last decade , due to its remarkable properties and potential applications@xcite .
the band structure of this exotic material is characterized by two distinct dirac points in the brillouin zone , dominating most of its physical results .
although there are a multitude of researches on graphene and other honeycomb lattices@xcite , the origin of the dirac points is still unclear . according to the von neumann - wigner theorem@xcite
, there must be some symmetry to protect the dirac points on the honeycomb lattice , while the robustness of dirac points during the deformation of the lattice structure@xcite excludes the possibility of a point group protection . as a result
, a novel symmetry is expected to be responsible for the dirac points . in this work
, we unveil the mysterious story behind the dirac points by showing that they are exactly protected by a kind of hidden symmetry on the lattice structure . as its name suggests , the hidden symmetry is not so obvious as usual symmetries such as the point group symmetry . in general
it can be described by a composite anti - unitary operator , which consists of a translation , a complex conjugation , and a sublattice exchange , and sometimes also a local gauge transformation and a rotation , or is the extension of the composite operator by a mapping method .
this kind of symmetry is seldom studied before and was firstly discovered by one of the authors in a toy model@xcite .
we find that the hidden symmetry on the honeycomb lattice evolves along with the variation of the parameters , which can perfectly explain the moving and merging of the dirac points and the quantum phase transition on the honeycomb lattice that have been theoretically predicted@xcite and experimentally observed@xcite .
* results * + * model*. to be specific , we consider the general honeycomb lattice as shown in fig.[fig1]a , where we define a bond angle @xmath0 as the angle between the bonds on the zigzag line and the horizontal direction .
when the bond angle takes peculiar values of @xmath1 and @xmath2 , the lattice reduces to the ideal honeycomb lattice , such as graphene , and to the brick - wall lattice as shown in fig.[fig1]b .
the general honeycomb lattice model can be well described by the bloch hamiltonian as ( the unit bond length is adopted ) @xmath3\sigma_x\nonumber\\ & + & [ 2t_1\cos(\cos\theta k_x)\sin(\sin\theta k_y)-t_2\sin k_y]\sigma_y\label{bhh}\end{aligned}\ ] ] where @xmath4 and @xmath5 denote the amplitudes of hopping as sketched in fig.[fig1]a ; @xmath6 and @xmath7 are the pauli matrices defined in the sublattice space . in order to find the hidden symmetry behind the the honeycomb lattice , an auxiliary square lattice with a hopping - accompanying @xmath8 phase is introduced as well , as shown in fig.[fig1]c , which has an intrinsic relation with the honeycomb lattice .
the bloch hamiltonian for the square lattice can be written as @xmath9 where @xmath10 and @xmath11 represent the amplitudes of hopping along the horizontal and vertical directions , respectively .
( for the derivations of the bloch hamiltonians of the honeycomb and square lattices see supplementary section s-1 . )
geometrically , the square lattice can transform into the ideal honeycomb lattice continuously in two steps .
first , the square lattice changes into the brick - wall lattice when the amplitude of hopping with a @xmath8 phase is tuned to zero , and then reaches the ideal honeycomb lattice by a deformation of the bond angle @xmath0 from 0 to @xmath12 , which can be understood with the help of fig.[fig1]a - c . besides the intuitive relation between these lattice structures in the real space , their band structures also strongly correlate to each other .
the energy bands are all characterized by two linear dirac cones in the brillouin zone as shown in fig.[fig1]d - f ( for the definitions of the brillouin zones see supplementary section s-2 ) .
more importantly , these dirac points are able to evolve continuously into each other with the variation of the lattice parameters . for the general honeycomb lattice with @xmath13
( @xmath14 is defined as the hopping amplitude ratio @xmath15 ) , the corresponding dirac points locate at @xmath16 in the brillouin zone . as a result , when the bond angle @xmath0 varies from @xmath12 to @xmath17 , the lattice first changes from the ideal honeycomb lattice into the brick - wall lattice , inducing a shift of the dirac points from @xmath18 to @xmath19 , as shown in fig.[fig1]d and fig.[fig1]e .
starting from the brick - wall lattice , the square lattice can be obtained by turning on the amplitude of hopping with a @xmath8 phase from @xmath17 to @xmath20 .
accordingly , the corresponding dirac points evolve from @xmath19 to @xmath21 , as shown in fig.[fig1]f .
it is impressive that during the whole evolution of the lattice , the dirac points are always stable without any gap opening .
we will show that this property can be perfectly explained by the protection of the hidden symmetry of the lattice structures .
* hidden symmetry and protection of dirac points*. firstly , we consider the auxiliary square lattice as shown in fig.[fig1]c .
one can verify that the square lattice is invariant under the action of the operator defined as @xmath22 where @xmath23 is a translation operator that moves the lattice along the horizontal direction by a unit vector @xmath24 ; @xmath25 is the complex conjugate operator ; @xmath6 is the pauli matrix representing the sublattice exchange ; @xmath26 is a local @xmath27 gauge transformation ( for details see supplementary section s-3 ) .
this kind of transformation invariance indicates a hidden symmetry of the square lattice@xcite .
it is easy to prove that the symmetry operator @xmath28 is anti - unitary , and its square is equal to @xmath29 .
mathematically , the hidden symmetry operator @xmath28 can be considered as a self - mapping of the square lattice model defined as @xmath30 where @xmath31 and @xmath32 are the bloch functions of the square lattice model . performing the hidden symmetry transformation on the bloch function leads to @xmath33 with @xmath34 ( methods ) .
if @xmath35 , where @xmath36 is a reciprocal lattice vector for the square lattice , then we can say that @xmath37 is a @xmath28-invariant point . in the brillouin zone ,
the @xmath28-invariant points are @xmath38 and @xmath39 .
after the hidden symmetry operator acts on the bloch function twice , we have @xmath40 .
therefore , the square of the hidden symmetry operator takes a value of @xmath41 in the bloch representation . for the @xmath28-invariant points
, we have @xmath42 at @xmath43 , while @xmath44 at @xmath45 .
since @xmath28 is an anti - unitary operator , we arrive at the important conclusion that the band energy must be degenerate at the @xmath28-invariant points @xmath43 , which are just the locations of the dirac points as shown in fig.[fig1]f . from the above discussion
, one can see that the dirac points on the auxiliary square lattice are exactly protected by the hidden symmetry @xmath28 .
( for details see supplementary section s-5a ) . in the following ,
we show that the hidden symmetry of the honeycomb lattice can be derived from that of the auxiliary square lattice .
we define a mapping from the honeycomb lattice model to the square lattice model as @xmath46 which depends on the bond angle @xmath0 and the hopping amplitude ratio @xmath14 , with @xmath47 being the bloch functions of the honeycomb lattice model .
the operator @xmath48 maps the bloch hamiltonian of the honeycomb lattice into that of the square lattice , which can be expressed as @xmath49 where the hopping amplitudes are related by @xmath50 and @xmath51 , and the wave vector @xmath52 is defined by @xmath37 as @xmath53 with @xmath54\}$ ] . performing the mapping procedure on the bloch function as well , one obtains @xmath55 .
( interpreting this mapping in an intuitive way see supplementary section s-4 ) . .
* b * , the case with @xmath56 . * c * , the case with @xmath57 . in the left panels ,
the yellow diamond areas represent the brillouin zone of the honeycomb lattice , which are equivalent to the areas enclosed the black solid lines ; the blue filled circles represent the @xmath58-invariant points @xmath59 . in the right panels ,
the square areas enclosed by the black solid lines represent the brillouin zone of the square lattice ; the blue and green filled circles represent the @xmath28-invariant points @xmath43 and @xmath45 , respectively ; the yellow areas are the image of the mapping @xmath48 for the brillouin zone of the honeycomb lattice .
the mapping @xmath48 concretely map the blue filled circles , the blue and red lines in left panels into the blue filled circles , the blue and red lines in the right panels , respectively . ] . *
b * , the case with @xmath56 . * c * , the case with @xmath57 . in the left panels ,
the yellow diamond areas represent the brillouin zone of the honeycomb lattice , which are equivalent to the areas enclosed the black solid lines ; the blue filled circles represent the @xmath58-invariant points @xmath59 . in the right panels ,
the square areas enclosed by the black solid lines represent the brillouin zone of the square lattice ; the blue and green filled circles represent the @xmath28-invariant points @xmath43 and @xmath45 , respectively ; the yellow areas are the image of the mapping @xmath48 for the brillouin zone of the honeycomb lattice .
the mapping @xmath48 concretely map the blue filled circles , the blue and red lines in left panels into the blue filled circles , the blue and red lines in the right panels , respectively . ] . *
b * , the case with @xmath56 . * c * , the case with @xmath57 . in the left panels ,
the yellow diamond areas represent the brillouin zone of the honeycomb lattice , which are equivalent to the areas enclosed the black solid lines ; the blue filled circles represent the @xmath58-invariant points @xmath59 . in the right panels ,
the square areas enclosed by the black solid lines represent the brillouin zone of the square lattice ; the blue and green filled circles represent the @xmath28-invariant points @xmath43 and @xmath45 , respectively ; the yellow areas are the image of the mapping @xmath48 for the brillouin zone of the honeycomb lattice .
the mapping @xmath48 concretely map the blue filled circles , the blue and red lines in left panels into the blue filled circles , the blue and red lines in the right panels , respectively . ] with the help of the mapping @xmath48 , we define an hidden symmetry operator of the honeycomb lattice as @xmath60 , which means the honeycomb lattice model is invariant under firstly a mapping into the square lattice , then a @xmath28-transformation , and finally an inverse mapping back to the honeycomb lattice .
therefore , the operation @xmath58 can be considered as a self - mapping of the honeycomb lattice model as @xmath61 applying this operator to the wave function , we obtain @xmath62 , where the final momentum is @xmath63 , with @xmath64 and @xmath65 being the shift of the wave vector @xmath37 due to the mapping @xmath48 . if @xmath66 with @xmath67 being the reciprocal lattice vector of the honeycomb lattice , then @xmath37 is a @xmath58-invariant point . for the honeycomb lattice model ,
the @xmath58-invariant points in the brillouin zone are @xmath68 .
it is straightforward to verify that @xmath69 , a direct action of which on the bloch function results in @xmath70}\psi_{h,\mathbf{k}}(\mathbf{r } ) $ ] .
substituting the @xmath58-invariant points @xmath59 into the above equation , we find @xmath71 .
since @xmath58 is an anti - unitary operator , there must exist band degeneracies at the @xmath72-invariant points @xmath59 .
( for details see supplementary section s-5b ) . in particular , when @xmath1 and @xmath13 , the @xmath73-protected degenerate points are @xmath74 , just the position of the dirac points on the ideal honeycomb lattice , such as graphene .
when @xmath2 and @xmath13 , the @xmath72-protected band degeneracies occur at @xmath75 , which correspond to the locations of the dirac points on the brick - wall lattice .
* explanation for the moving and merging of dirac points and the quantum phase transition*. we have proved above that the dirac points on the honeycomb lattice are protected by the hidden symmetry @xmath58 . more generally , the moving and merging of dirac points on the honeycomb lattice , which has been predicted theoretically@xcite and observed experimentally@xcite , can also be explained by the hidden symmetry . since the hidden symmetry operator @xmath58 contains the parameters @xmath0 and @xmath14 , the locations of the @xmath58-protected dirac points , @xmath76 , are also functions of the two parameters .
as the hopping amplitude ratio @xmath14 starts to increase from 1 , the dirac points move away from each other . when @xmath14 reaches @xmath77 , two dirac points merge into a single one at the corner of the brillouin zone .
if @xmath14 increases further , there is no solution to the @xmath72-invariant points , thereby the dirac points vanish , with a gap opening simultaneously . as a result , @xmath56 is the critical point of the quantum phase transition .
we can interpret the above conclusion in a more intuitive way by mapping the brillouin zone of the honeycomb lattice into that of the square lattice , as shown in fig.[fig2 ] .
it turns out that such a mapping is not surjective , which means that the image of the brillouin zone of the honeycomb lattice is part of the brillouin zone of the square lattice . in the parameter interval of @xmath78 ,
the image covers the @xmath28-protected degenerate points @xmath43 in the brillouin zone of the square lattice , as shown in fig.[fig2]a .
thus , there always exist two points @xmath59 in the brillouin zone of the honeycomb lattice mapping into the @xmath28-protected degenerate points @xmath43 in the brillouin zone of the square lattice . when @xmath56 , the two equivalent points locating at the corners of the brillouin zone of the honeycomb lattice map into the @xmath28-protected degenerate points @xmath43 , as shown in fig.[fig2]b
. therefore , the dirac points on the honeycomb lattice merge .
when @xmath79 , the image of the brillouin zone of the honeycomb lattice can not cover the @xmath28-protected degenerate points @xmath43 , as shown in fig.[fig2]c .
therefore , there is no point in the brillouin zone of the honeycomb lattice mapping into the @xmath28-protected degenerate points @xmath43 .
thus , the dirac points disappear and a gap opens . * discussion*. in summary
, we have found a hidden symmetry on the honeycomb lattice and proved that the hidden symmetry protects the dirac points on the honeycomb lattice .
the hidden symmetry evolves along with the parameters , such as the bond angle @xmath0 and the hopping amplitude ratio @xmath14 , which provides a perfect explanation on the moving and merging of the dirac points and the quantum phase transition on the honeycomb lattice .
our research unfolds a new perspective on the symmetry protected band degeneracy , which is totally different from the conventional ones , such as the band degeneracy protected by point groups or time reversal symmetry .
such novel hidden symmetry can greatly enrich and deepen our understanding of the band degeneracy , which will have important applications in modern condensed matter physics , especially , in the topics of dirac ( weyl ) semimetals and other topological semimetals . * methods * * the transformation of the wave vectors under the action of the operator @xmath28 . * the square lattice is invariant under the action of the operator @xmath28 , which is anti - unitary .
we suppose that the bloch functions of the square lattice model have the form as @xmath80 where @xmath81 with @xmath82 .
then , the hidden symmetry operator @xmath28 acts on the bloch functions as follows @xmath83 } \label{bloch2}\end{aligned}\ ] ] because @xmath28 is the symmetry operator for the square lattice , @xmath84 must be a bloch function of the square lattice model .
comparing equation ( [ bloch2 ] ) with equation ( [ bloch ] ) , we have @xmath85 which can be regarded as the transformation of the wave vector under the action of the operator @xmath28
. 99 novoselov , k.s .
_ et al . _
electric field fffect in atomically thin carbon films .
_ science _ * 306 * , 666 - 669 ( 2004 ) .
_ _ _ _ novoselov , k.s .
_ et al .
_ two - dimensional gas of massless dirac fermions in graphene .
_ nature _ * 438 * , 197 - 200 ( 2005 ) .
_ _ _ _ zhang , y. , tan , y .- w . ,
stormer , h.l . & kim , p. experimental observation of the quantum hall effect and berry s phase in graphene
. _ nature _ * 438 * , 201 - 204 ( 2005 ) . _ _
castro neto , a.h . ,
guinea , f. , peres , n.m.r . ,
novoselov , k.s . & geim , a.k .
the electronic properties of graphene . _ rev .
mod .
phys . _ * 81 * , 109 - 162 ( 2009 ) .
_ _ wallace , p.r .
the band theory of graphite .
_ phys .
rev . _ * 71 * , 622 - 634 ( 1947 ) . _ _
li , g. & andrei , e.y .
observation of landau levels of dirac fermions in graphite .
_ nature phys . _ * 3 * , 623 - 627 ( 2007 ) .
_ _ park , c .- h .
, yang , l. , son , y .- w . , cohen , m.l . & louie , s.g .
anisotropic behaviours of massless dirac fermions in graphene under periodic potential .
_ nature phys . _ * 4 * , 213 - 217 ( 2008 ) . _ _ park , c .- h . & louie , s.g .
making massless dirac fermions from a patterned two - dimensional electron gas .
_ nano lett . _
* 9 * 1793 - 1797 ( 2009 ) . _
_ gibertini , m. _ et al . _ enginering artificaial graphene in a two - dimensional electron gas . _
phys .
rev .
b _ * 79 * , 241406 ( 2009 ) . _
_ _ _ asano , k. & hotta , c. designing dirac points in two - dimensional lattices .
_ phys .
rev .
b _ * 83 * , 245125 ( 2011 ) . _ _ gomes , k.k . , mar , w. , ko , w. , guinea , f. & manoharan , h.c .
designer dirac fermions and topological phases in molecular graphene .
_ nature _ * 483 * , 306 - 310 ( 2012 ) . _ _ de gail , r. , fuchs , j .- n . , goerbig , m.o . ,
pichon f. & montambaux , g. manipulation of dirac points in graphene - like crystals _ physica b _ * 407 * , 1948 - 1952 ( 2012 ) . _
_ bellec , m. , kuhl , u. , montambaux , g. & mortessagne , f. topological transition of dirac points in a microwave experiment .
_ phys .
rev .
lett . _
* 110 * , 033902(2013 ) . _
_ von neumann j. & wigner e. on the behavior of eigenvalues in adiabatic processes .
_ z. phys . _ * 30 * , 467 - 470 ( 1929 ) .
_ _ landau , l.d .
& lifshitz e. _ quantum mechanics : non - relativistic theory 3rd ed .
_ , ( butterworth - heinemann , 1981 ) . _ _ tarruell , l. , greif , d. , uehlinger , t. & esslinger , t. creating , moving and merging dirac points with a fermi gas in a tunable honeycomb lattice .
_ nature _ * 483 * , 302 - 305 ( 2012 ) . _
_ hou , j .-
m .
hidden - symmetry - protected topological semimetals on a square lattice .
_ phys .
rev .
lett . _ * 111 * , 130403 ( 2013 ) .
_ _ hou , j .-
m . moving and merging of dirac points on a square lattice and hidden symmetry protection .
_ phys .
rev .
b _ * 89 * , 235405 ( 2014 ) . _ _ hasegawa , y. , konno , r. , nakano , h. & kohmoto , m. zero modes of tight - binding electrons on the honeycomb lattice . _
phys .
rev .
b _ * 74 * , 033413 ( 2006 ) . _ _ zhu , s .- l . , wang , b. & duan , l .- m .
simulation and dectection of dirac fermions with cold atoms in an optical lattice .
_ phys .
rev .
lett . _ * 98 * , 260402 ( 2007 ) . _ _ wunsch , b. , guinea , f. & sols , f. dirac - point engineering and topological phase transitions in honeycomb optical lattices .
_ new j. phys . _ * 10 * , 103027 ( 2008 ) . _ _ pereira , v.m . ,
castro neto , a.h .
& peres , n.m.r .
tight - binding approach to unianixal strain in graphene .
_ phys .
rev .
b _ * 80 * , 045401 ( 2009 ) . _ _ montambaux , g. , pichon , f. , fuchs , j .- n . &
goerbig , m.o .
mergering of dirac points in a two - dimensional crystal .
_ phys .
rev .
b _ * 80 * , 153412 ( 2009 ) . _ _ |
we begin by presenting some more details concerning our semiclassical evaluation of the disorder correlation function , @xmath10 , and corresponding susceptibility @xmath21 .
we consider non - interacting electrons in a weak , perpendicular magnetic field . in terms of retarded and advanced single particle green functions , @xmath36 , @xmath10
may be written as @xmath37 , where @xmath38 implies the inclusion of connected diagrams only . using a diagrammatic approach introduced by altland and gefen @xcite
the field sensitive part of @xmath10 can be expressed as a sum including cooperon type diagrams @xmath39 defined by @xmath40^n = \int \prod_{j=1}^{n } d^dr_j\prod_{m=1}^{n } \zeta^{(c ) } ( { \bf r}_m,{\bf r}_{m+1};\omega ; h ) \quad ; \quad { \bf r}_{n+1 } \equiv { \bf r}_1 \ ; . \label{sn}\ ] ] here @xmath41 , @xmath42 is the disorder averaged single particle green function , @xmath43 , and @xmath44 .
semiclassically , @xmath45 can be expressed as a sum over classical trajectories @xmath46 between @xmath47 and @xmath48 @xcite , @xmath49 } \ ; .\ ] ] the prefactor @xmath50 includes the classical phase space density , @xmath51 stands for the classical action along an orbit @xmath46 ( in the absence of disorder ) including the maslov index , and @xmath52 is the orbit length .
@xmath53 is then given in terms of pairs of classical paths which explicitly include the effect of boundary scattering .
however most pairs ( of different paths ) produce oscillating contributions which we assume to vanish after energy or size averaging .
the main contribution to the field sensitive part of @xmath20 arises from diagonal terms ( otherwise known as the cooperon channel ) obtained by pairing paths with their time reverse . assuming that the magnetic field affects the phase of the particles but not their trajectories we can write @xmath54 where @xmath55 \ ; . \label{zsc}\ ] ] here , @xmath56 is the period of the trajectory , @xmath57 is the vector potential , @xmath58 , and the level broadening was introduced via @xmath59 .
( [ zsc ] ) depends , besides @xmath0 , only on the system without disorder and holds for both integrable and chaotic geometries .
the disorder induced contribution to the average susceptibility is given by ( @xmath60 ) : @xmath61 where the bulk landau susceptibility is @xmath62 for spinless electrons . in eq .
( [ stos ] ) the @xmath63 are now assumed to contain diagonal terms only and their contributions ( eq .
( [ sn ] ) ) can be calculated by diagonalising @xmath64 which in general can not be done analytically .
however we can use the fact that all the variations of @xmath64 occur on classical lengthscales ; rapid oscillations on the scale of @xmath65 cancel out .
it is thus possible to discretise the `` classical '' operator @xmath64 on a lattice in space with grid size greater than @xmath65 . by summing over trajectories between lattice cells one
can compute @xmath64 , and thereby @xmath66 , efficiently by numerical means@xcite .
the propagator @xmath64 ( above eq .
( [ zsc ] ) ) is made up of a summation over all diagonal pairs of paths ( including boundary scattering ) between any two given impurities situated at @xmath47 and @xmath48 .
on taking the trace over @xmath67 propagators @xmath64 , one sees that the field sensitive part of @xmath68 , eq .
( [ sn ] ) , consists of a summation over flux - enclosing closed pairs of paths ( in position space ) involving @xmath67 impurities and an arbitrary number of boundary scattering events .
however this summation does not include closed pairs of paths which follow periodic orbits of the corresponding clean system .
such paths involve zero momentum transfer between the green functions at the impurity positions ; they actually represent disconnected diagrams which are included in @xmath22 and must not be counted again in the determination of @xmath21 .
it is the presence of these periodic orbits in the determination of @xmath22 that leads to strong sensitivity with respect to the system geometry @xcite ( see below ) . in the following
we apply the above formalism to the case of an ensemble of disordered square billiards .
gefen , braun , and montambaux ( gbm ) @xcite considered the contribution of trajectories longer than @xmath0 to @xmath21 in an approximate way , while richter , ullmo , and jalabert ( ruj ) @xcite calculated @xmath22 for a square by assuming that the disorder perturbs the phase , but not the trajectory , of semiclassical paths of the corresponding clean geometry .
we first compute the @xmath39 ( eq . ( [ sn ] ) ) for the square geometry by employing the extended zone scheme@xcite to write @xmath69 as a sum of propagators along straight line paths .
we then perform a complete calculation of @xmath21 , eq .
( [ stos ] ) , in the elastic and inelastic regimes and compare the results with those by gbm and ruj .
= 0.5 = 0.52 [ elastic ]
fig . 1 shows @xmath71 as a function of @xmath0 for a typical experimental value of @xmath72 .
the lower curve is for @xmath73 ( i.e. @xmath74 at @xmath72 ) and the upper is for @xmath75 @xcite .
note that in the diffusive regime , @xmath32 , there is linear increase with @xmath0 in agreement with ref .
@xcite . for @xmath31
, we find a weak dependence of @xmath71 on @xmath0 .
our result is on the whole close to the prediction by gbm @xcite who found a paramagnetic @xmath0-independent contribution , @xmath76 , shown as the dashed horizontal line in fig . 1 for @xmath73 .
as @xmath0 increases further , @xmath71 decreases ; we discuss this behaviour in more detail later when considering the inelastic regime .
the inset of fig .
1 shows @xmath71 as a function of @xmath77 for @xmath73 . from the top ,
the five curves are for values of @xmath78 , @xmath79 , @xmath80 , @xmath81 and @xmath82 . for all disorder strengths , @xmath71 is paramagnetic and it increases linearly with @xmath77 . for @xmath83
the gradient of the curves is approximately independent of @xmath0 and we find it to be @xmath84 . however there is a @xmath77 independent offset to the curves which is @xmath0 dependent and not described by gbm .
fig . 2 shows @xmath71 as a function of @xmath0 for @xmath72 and @xmath86 ( i.e. @xmath87 ) .
circular points correspond to our full numerical results , while triangular points represent the contribution of @xmath88 only ( minus the disconnected part ) in eq .
( [ stos ] ) .
the solid line is the equation @xmath89 where @xmath90 and @xmath91 are fitting parameters .
the dashed line displays an equation of the same type but with @xmath92 and @xmath90 as the only fitting parameter . for @xmath93
the susceptibility is dominated by the contribution with the lowest number of impurity scatterings @xmath88 . as @xmath94 is reduced , progressively more terms in the summation of eq .
( [ stos ] ) become relevant , and there is good agreement with the solid line fit for @xmath95 . the inset of fig
. 2 shows @xmath71 as a function of @xmath77 for @xmath86 . from the top ,
the three curves are for values of @xmath78 , @xmath79 and @xmath96 .
the symbols correspond to our results and the solid lines to @xmath97 $ ] where @xmath98 and @xmath99 are fitting parameters .
= 0.48 = 0.48 [ compare ]
we compare the magnitude of @xmath71 with that of @xmath100 for squares .
it has been shown @xcite that the low field susceptibility of an ensemble of _ clean _ squares is dominated by the shortest flux enclosing periodic orbits of length @xmath101 and their repetitions over a broad range of temperature ( and thus inelastic scattering strengths ) . for the ballistic white noise case considered here ,
the effect of disorder averaging on the susceptibility was described by an additional damping @xmath102 of the response of the clean system @xcite .
this result corresponds to @xmath100 including the disorder damping @xmath103 of the single particle green functions .
we use the results of ruj @xcite at zero temperature and introduce the level broadening @xmath27 in the same way as for the disorder correlations above @xcite .
it is then possible to sum the contribution of all repetitions of the fundamental orbit explicitly which gives @xmath104 $ ] .
3 shows this expression for @xmath100 ( dashed line ) and our numerical results for @xmath71 ( solid line ) as a function of @xmath105 for @xmath72 and @xmath106 .
although @xmath107 for small @xmath108 and vice versa for large @xmath108 , it is clear that both contributions are relevant over a broad range of @xmath108 .
we make an estimate for the value of @xmath108 at which the contributions of @xmath100 and @xmath71 are equal by comparing the above analytic approximation with that given by gbm .
we find for @xmath109 that @xmath21 is larger than @xmath22 for @xmath26 greater than a crossover value , @xmath110 , given by @xmath111^{-1 } \ ; . \label{cgam}\ ] ] the inset of fig .
3 shows this estimate ( solid line ) compared to points ( circles ) obtained by comparing our numerical results for @xmath71 with the analytic expression for @xmath100 .
the experiment@xcite on the orbital magnetism of ensembles of squares had estimated values for the elastic mean free path of @xmath112 , for the phase - coherence length of @xmath113 and for the thermal cutoff length of @xmath114 .
hence , the lengthscale @xmath26 ( eq . ( [ lphi ] ) ) is determined by the shorter length @xmath115 .
3 shows that for these experimental parameters ( and for white noise disorder ) both the disorder and size - induced correlations are relevant , however the latter contribution is dominant .
the measured value of the susceptibility at _ low _ temperature was @xmath116 , with an uncertainty of about a factor of four . after including a spin factor of 2 , the combined contributions @xmath117 and @xmath118 calculated above , together with an interaction contribution of the same order @xcite , are in broad agreement with the experimental result .
we note , however , that a theoretical explanation of the temperature dependence of the measured susceptibility is still lacking . experimental ballistic structures as those in ref .
@xcite are usually characterised by smooth disorder potentials .
the effect of smooth disorder on @xmath119 has been analysed in ref . @xcite showing that the reduction of the clean contribution is less strong than for white noise disorder and no longer exponential .
smooth disorder effects can be incorporated into the present calculation by introducing an angle - dependent cross section for the impurity scattering between two trajectory segments .
for systems with a generic chaotic , clean counterpart we obtain an analytical estimate for @xmath71 in the elastic regime after transforming the sum over densities @xmath120 in eq .
( [ zsc ] ) into probabilities @xmath121 to propagate classically from @xmath122 to @xmath123 at time @xmath46 accumulating an `` area '' @xmath124 @xcite . assuming a gaussian `` area '' distribution with a variance @xmath125 , which is taken to be @xmath0 independent , we find @xmath126 summation of the @xmath39 , eq .
( [ stos ] ) , leads to @xmath127 @xmath71 is shown as the full line in fig . 4 which is rather similar to the corresponding curve in fig . 3 for the square geometry
a corresponding approximation for @xmath128 @xcite is shown as the dashed line in fig .
4 . both can be shown to add up to an @xmath0-independent response @xmath129 for generic chaotic geometries .
disorder - induced spectral correlations and their effect on the magnetic susceptibility in the non - diffusive regime @xmath2 were considered .
we focused on the square billiard , whose corresponding clean geometry is integrable , and showed that there are two distinct regimes of behaviour depending on the relative magnitudes of @xmath0 and @xmath24 .
this approach enabled us to study the complete crossover from diffusive to clean systems for arbitrary values of @xmath0 and @xmath24 ( smaller than @xmath33 ) .
note that it may be possible to calculate the susceptibility for values of @xmath24 and @xmath0 greater than @xmath33 using a non - perturbative approach such as the ballistic @xmath125 model @xcite .
we are grateful to y. gefen , s. kettemann , d. e. khmelnitskii , m. leadbeater , i. v. lerner and p. walker for useful discussions .
we thank the isaac newton institute for mathematical sciences , cambridge , where part of this research was performed . 99 . . . . .
the spectral form factor for ballistic systems with rigid disks ( spheres ) as impurities was studied recently by . . .
preprint , to be submitted to _ phys .
by comparison with random matrix results @xmath75 has been shown to correspond to infinite @xmath24 ; . .
@xcite the orbits are damped by including temperature smearing rather than level broadening . | we study disorder - induced spectral correlations and their effect on the magnetic susceptibility of mesoscopic quantum systems in the non - diffusive regime . by combining a diagrammatic perturbative approach with semiclassical techniques we perform impurity averaging for non - translational invariant systems .
this allows us to study the crossover from clean to diffusive systems . as an application we consider the susceptibility of non - interacting electrons in a ballistic microstructure in the presence of weak disorder .
we present numerical results for a square billiard and approximate analytic results for generic chaotic geometries .
we show that for the elastic mean free path @xmath0 larger than the system size , there are two distinct regimes of behaviour depending on the relative magnitudes of @xmath0 and an inelastic scattering length .
euromacr phase - coherent , disordered conductors where the electron motion equals a random walk between impurities have traditionally been of interest in mesoscopic physics @xcite .
random walks occur in the diffusive regime where the elastic mean free path @xmath0 is much smaller than the system size @xmath1 . on the other hand ,
the development of high - mobility semiconductor heterostructures , combined with advanced lithographic techniques , have allowed the confinement of electrons to two - dimensional microstructures of controllable , non - random geometry .
they have been coined `` ballistic '' since @xmath2 .
nevertheless , residual impurity scattering is nearly unavoidable even in these systems and it became clear that disorder can be strong enough to mix energy levels and effect the two - level correlation function , @xmath3 , even in the `` ballistic '' regime @xcite .
there it is necessary to consider both disorder averaging , @xmath4 , and size ( or energy ) averaging , @xmath5 .
after such averaging the two - level correlation function may be divided into two separate terms @xcite , @xmath6 and @xmath7 , where @xmath8 denotes the single particle density of states and @xmath9 its mean part .
@xmath10 is a measure of _ disorder - induced _ correlations of @xmath8 , while @xmath11 is given by _ size - induced _ correlations .
the orbital magnetism of isolated mesoscopic systems has been the subject of much theoretical interest , in particular because it is sensitive to spectral correlations : for a system with a fixed number of particles it is necessary to consider averaging under canonical conditions @xcite resulting in a large contribution to the average magnetism .
the corresponding susceptibility is given by @xcite @xmath12 here @xmath13 is the magnetic field , @xmath14 is the mean level spacing and @xmath15 is the variance in the number of energy levels within an energy interval of width equal to the chemical potential , @xmath16 .
this variance is related to @xmath17 by integration of the level energies @xmath18 , @xmath19 over the energy interval . in the following we will label the contributions to the susceptibility ,
corresponding to @xmath20 and @xmath11 , as @xmath21 and @xmath22 , respectively , so that @xmath23 for real systems , besides @xmath0 , there are additional relevant lengthscales at which inelastic scattering ( @xmath24 ) or temperature smearing ( @xmath25 ) produce a damping of propagation .
for clarity we refer to such a lengthscale as @xmath24 , although we assume that similar general arguments will hold for finite @xmath25 . for ballistic motion ,
@xmath26 is related to the level broadening @xmath27 by @xmath28 with @xmath29 being the fermi momentum .
it divides the `` ballistic '' regime into two sub - regimes . in the first , @xmath30
, the particle motion is nearly ballistic since damping due to inelastic scattering typically occurs before impurity scattering ; for the remainder of this paper we refer to this regime as _
inelastic_. in the second , @xmath31 , a particle may scatter many times off impurities before scattering inelastically and we refer to this regime as _
elastic_. in this paper we employ a systematic semiclassical approach to calculate the contribution of orbits _ of all lengths _ in the elastic and inelastic regimes to weak field @xmath21 for microstructures with white noise disorder @xcite .
energy and impurity averaging are performed within a diagrammatic perturbation approach applicable to the `` ballistic '' regime .
in contrast to many techniques valid for bulk disordered systems , @xmath32 , we do not assume translational invariance .
instead , we use an approach related to the `` method of trajectories '' devised for thin superconducting films @xcite and we write green functions semiclassically in terms of classical paths which include the effect of boundary scattering . note that similar @xcite , and alternative @xcite ,
methods have been used to consider weak localization in thin films in a parallel magnetic field .
our approach allows us to study the complete crossover from diffusive to clean systems for arbitrary values of @xmath24 ( smaller than @xmath33 , where @xmath34 is the heisenberg time ) .
we present results for ballistic systems with both chaotic and integrable dynamics in the clean limit . as an example of an integrable geometry
, we treat the case of the square billiard .
experiments @xcite on the orbital magnetism of ensembles of squares were performed in the `` ballistic '' ( inelastic ) regime , motivating theoretical studies of the susceptibility for @xmath35 @xcite . |
we consider here spin - half nuclei , quantum bits ( qubits ) , whose computation - basis states align in a magnetic field either in the direction of the field ( @xmath0 ) , or in the opposite direction ( @xmath1 ) . several such bits ( of a single molecule ) represent a binary string , or a register . a macroscopic number of such registers / molecules can be manipulated in parallel , as done , for instance , in _ nuclear magnetic resonance ( nmr)_. from the perspective of _ nmr quantum computation ( nmrqc ) _ @xcite ( for a recent review see @xcite ) , the spectrometer that monitors and manipulates these qubits / spins can be considered a simple `` quantum computing '' device ( see for instance refs .
@xcite and @xcite and references therein ) .
the operations ( gates , measurements ) are performed in parallel on many registers .
the probabilities of a spin to be up or down are @xmath2 and @xmath3 , where @xmath4 is called the _ polarization bias_. , where @xmath5 is the boltzmann constant , @xmath6 is the gyromagnetic coefficient , @xmath7 is the intensity of the magnetic field and @xmath8 is the temperature of the heat bath . for small values ,
@xmath9 at equilibrium at room temperature , nuclear - spin polarization biases are very small ; in common nmr spectrometers @xmath4 is at most @xmath10 .
while the polarization bias may be increased by physical cooling of the environment , this approach is very limited for liquid - state nmr , especially for in - vivo spectroscopy . to improve polarization , including selective enhancement , various `` effective cooling '' methods increase the polarization of one or more spins in a transient manner ; the cooled spins re - heat to their equilibrium polarization as a result of thermalization with the environment , a process which has a characteristic time of @xmath11 .
an interesting effective cooling method is the reversible ( or unitary ) entropy manipulation of srensen and schulman - vazirani @xcite , in which data compression tools and algorithms are used to compress entropy onto some spins in order to cool others .
such methods are limited by shannon s entropy bound , stating that the total entropy of a closed system can not decrease .
the reversible cooling of schulman - vazirani was suggested as a new method to present scalable nmrqc devices .
the same is true for the algorithmic cooling approach @xcite that is considered here .
_ algorithmic cooling ( ac ) _ combines the reversible effective - cooling described above with thermalization of spins at different rates ( @xmath11values ) @xcite .
ac employs slow - thermalizing spins ( _ computation spins _ ) and rapidly thermalizing spins ( _ reset spins _ ) .
alternation of data compression steps that put high entropy ( heat ) on reset spins , with thermalization of hot reset spins ( to remove their excess entropy ) can reduce the total entropy of suitable spin systems far beyond shannon s bound .
let us describe in detail the three basic operations of ac . in some algorithms ,
the reset spins can be used directly in the compression step @xcite .
these algorithms have already led to some experimental suggestions and implementations @xcite . an efficient and experimentally feasible ac technique , termed
_ `` practicable algorithmic cooling ( pac ) '' _
@xcite , combines _ polarization transfer ( pt ) _ , reset , and rpc among three spins called _ 3-bit - compression ( 3b - comp)_. the subroutine _ 3b - comp _ may be implemented in several ways .
one implementation is as follows @xcite : + _ exchange the states @xmath12 .
+ leave the other six states ( @xmath13 , @xmath14 , etc . ) unchanged . _
+ if 3b - comp is applied to three spins \{c , b , a } with biases @xmath15 , then spin @xmath16 will acquire the bias @xcite : @xmath17 pac1 and pac2 consider _ purification levels _ , where spins are cooled by a factor of @xmath18 at each successive level ; at the @xmath19 purification level , @xmath20 ) with biases that are @xmath21 , therefore calculations are usually done to leading order in @xmath22 . ] .
the practicable algorithm pac2 @xcite uses any odd number of spins , @xmath23 : one reset spin and @xmath24 computation spins ( see appendix [ app : alg - pac ] for a formal description of pac2 ) .
pac2 cools the spins such that the coldest spin can , ideally , attain a bias of @xmath25 .
when all spins are cooled , the following biases are reached : @xmath26 this proves an exponential advantage over the best possible reversible cooling techniques ( e.g. , of refs .
@xcite and @xcite ) , which are limited to a cooling factor of @xmath27 . as pac2 can be applied to as few as 3 spins ( @xmath28 ) to 9 spins ( @xmath29 ) , using nmrqc tools , it is potentially suitable for near future applications @xcite .
pac is simple , as all compressions are applied to three spins and use the same subroutine , _
3b - comp_. pac2 always applies compression steps ( _ 3b - comp _ ) to three identical biases . in general , for three spins with biases , @xmath30 , where all biases @xmath31 , 3-bit compression would attribute spin @xmath16 with the bias @xcite : @xmath32 using eq .
[ eq:3b - comp ] , several algorithms were designed , including pac3 , which always applies 3b - comp to bias configurations of the form @xmath33 ( see formal description in appendix [ app : alg - pac ] ) .
when all spins are cooled , the following biases are obtained : @xmath34 the framework of pac was extended to include multiple cycles at each recursive level .
consider the simplest case , whereby a three - spin system @xmath35 with equal bias @xmath4 , is cooled by repeating the following procedure @xmath36 times : 1 . _
3b - comp _ on @xmath35 increasing the bias of @xmath16 .
2 . reset the biases of @xmath37 and @xmath38 back to @xmath4 .
the bias of spin c increases to @xmath39 .
thus the biases @xmath40 are asymptotically obtained for the three spins , as noted first by fernandez @xcite .
this variant , where @xmath41 , is the first exhaustive spin - cooling algorithm . by generalizing this `` fernandez algorithm '' to more spins
, we obtained @xcite a bias configuration that asymptotically approaches the fibonacci series ; when all spins are cooled , the following biases are reached : @xmath42 algorithms based on the 3b - comp were recently reviewed @xcite . following the fibonacci algorithm ,
a related algorithm was described , which involves compression on four adjacent spins @xcite ; the tribonacci algorithm reaches biases in accord with the tribonacci series ( also called 3-step fibonacci series ) @xmath43 , where each term is the sum of the three previous terms .
similarly , one can obtain general @xmath44-step fibonacci series , where each term is the sum of the previous @xmath44 elements , respectively .
summing over all previous terms produces the geometric sequence with powers of two : @xmath45 the ac that reaches this series was termed @xcite , `` all - bonacci ac '' .
we conclude that , with the same number of spins , exhaustive ac reaches a greater degree of cooling than pac algorithms . for example , the cooling factor achieved by fibonacci with 9 spins is @xmath46 .
the corresponding enhancement for pac2 is @xmath47 , and for pac3 it is about @xmath48 . the _ partner pairing algorithm ( ppa ) _ was shown to achieve a superior bias than all previous cooling algorithms , and was proven to be optimal @xcite .
the all - bonacci algorithm appears to produce the same cooling as the ppa : @xmath49 ( this was verified numerically for small biases @xcite ) . in all - bonacci , with @xmath50 spins , the coldest spin approaches a bias of @xmath51 ( as long as @xmath52 ) . for the ppa
, a calculation of the cooling that yields @xmath53 was not performed , yet it was proven that @xmath54 .
see @xcite for a proof and see @xcite for calculations in the case of @xmath52 .
the focus of this paper is on simple algorithms that are semi - optimal and also practicable , owing to their exclusive reliance on simple logic gates .
such algorithms could be implemented in the lab if proper physical conditions ( e.g. , @xmath11 ratios ) are achieved , and thus they could have practical implications in the near future .
in contrast , the optimal algorithms mentioned above , the ppa and the all - bonacci , do not belong to this category .
the all - bonacci requires an unreasonable number of reset steps , and the permutations required in the sorting step of the ppa were not yet translated into simple 2-spin and 3-spin gates .
the next section describes @xmath55pac , a new semi - optimal algorithm based on pac , and its special variant , mpac , which cool @xmath23 spins such that the coldest spin is asymptotically cooled by a factor of @xmath56 .
the optimal algorithms , the ppa and all - bonacci , need about half the spins ( more precisely , @xmath57 spins ) in order to cool ( asymptotically ) the coldest spin to @xmath58 ; we thus see that mpac is semi - optimal in the sense that it needs twice as much spins to reach nearly the same optimal bias .
section [ sec : fib - variants ] discusses simple variants of the fibonacci algorithm : @xmath59-fibonacci @xcite , and mfib , which fixes the number of cycles at each recursive level .
section [ sec : comparison ] compares the new algorithms , mpac and mfib , to previous cooling algorithms , including the ppa , fibonacci , all - bonacci , and pac algorithms .
we show that practicable versions of mpac and mfib ( with small @xmath36 ) reach significant cooling levels for 5 - 11 spin systems ; moreover , in the case of a 5-spin system , semi - optimal cooling levels ( half the optimal cooling ) are attained with reasonable run - time .
section [ sec : pure - qubits ] provides some analysis of sopac in case one would like to purify qubits as much as needed in order to obtain scalable nmrqc .
we examine the resources required to obtain a highly polarized spin .
[ sec : mpac - variants ] the fernandez and fibonacci algorithms described above illustrate the improved level of cooling attainable ( asymptotically ) by repeated compressions that involve the target spin .
practicable cooling algorithms that perform a small number of such cycles ( at each recursive level ) offer reasonable cooling at a reasonable run - time . in this section
we describe mpac , a new pac - based algorithm , which approaches nearly half the optimal cooling level with only a small number of cycles .
the mpac algorithm that we now present is a generalization of pac2 ( see appendix [ app : alg - pac ] for details on pac2 ) .
@xmath60 is the purification level , @xmath61 is the bit index , @xmath62 stands for 3b - comp on spins @xmath63 which increases the bias of @xmath44 , @xmath64 denotes a polarization transfer from bit @xmath37 to bit @xmath38 ( or for simplicity , just a swap between their states ) , and @xmath65 is a reset , setting the bias of the reset spin , @xmath66 , to @xmath4 .
the algorithm only uses elementary gates : a single gate operates either on a single spin or on a pair of spins ( e.g. , pt between adjacent spins ) , or on three spins ( 3b - comp and pt between next - to - nearest neighbors ) .
@xmath67 takes @xmath61 spins at equilibrium and attributes bit @xmath44 with a bias of @xmath68 .
_ mpac : _ + @xmath69^{m } pt(k-2\rightarrow k)\ ; m_{j-1}(k-2 ) .
\nonumber\end{aligned}\ ] ] with three spins @xmath70^{m}pt(1\rightarrow 3)m_0(1).\end{aligned}\ ] ] with five spins @xmath71^{m}pt(3\rightarrow 5)m_1(3).\ ] ] the recursive formulas above are written from right to left , such that the first step of @xmath67 is reset of the reset spin , @xmath65 , followed by pt from spin 1 to spin 3 , and @xmath36 repetitions of the 4-step sequence in square brackets that ends with _ 3b - comp_. notice that the number of spins required by mpac to achieve a purification level @xmath72 is the same as for pac2 : @xmath73 but @xmath68 now depends on the number of cycles , @xmath36 : @xmath74 .
for @xmath75 we get back the algorithm pac2 , where @xmath76 , thus 1pac @xmath77 pac2 ; we often retain here the original name , pac2 .
asymptotically , for @xmath78 we get @xmath79 . for three spins , @xmath80 and @xmath81
this is the fernandez algorithm described above , and for sufficiently large @xmath36 , spin @xmath16 will acquire a final bias of @xmath82 @xcite .
this is calculated via @xmath83 .
similarly , for five spins and sufficiently large @xmath36 , the final bias of the @xmath84 spin is @xmath85 .
the choice of @xmath36 has a strong influence on the run - time , but fortunately , the polarization enhancement also increases rapidly with @xmath36 . for small spin - systems ( up to about 10 spins ) , very small values of @xmath36 ( 26 )
are sufficient .
figure [ fig : mpac - analysis - for - j<7 ] compares the cooling factors obtained by such mpac variants up to 13 spins .
notably , 6pac cools to a similar extent as @xmath86-pac .
it is also evident that 2pac cools significantly better than the single - cycle variant ( 1pac @xmath77 pac2 ) .
, where @xmath87 ) for small spin - systems after mpac with various @xmath36 ( see eq .
[ eq : mpac - definition ] ) . ] the run - time of mpac ( neglecting @xmath88 and _ 3b - comp _ steps ) is @xmath89 in reset - time units with @xmath90 . the run - time is therefore reduced by a factor of @xmath91 . ] . if @xmath75 ( mpac is pac2 ) the run - time @xcite is @xmath92 . in general , for any integer @xmath93 @xmath94 for instance , for 21 spins , @xmath95 , @xmath96 , we get @xmath97 ( as long as the final bias is still much smaller than 1 ) , and for the same @xmath98 , but @xmath99 , @xmath100 . for 11 spins , @xmath101 , and @xmath99 yields @xmath102 in general , for @xmath103 and sufficiently large @xmath36 , spin @xmath61 is attributed a bias of @xmath104 .
we denote the asymptotic case ( @xmath78 ) by @xmath105pac .
we now generalize mpac by replacing the constant @xmath36 by a vector @xmath106 .
this added flexibility , in which a different value of @xmath36 is associated with each cooling level , was found to be beneficial in the analysis of the fibonacci algorithm @xcite , as we explain in section [ sec : fib - variants ] .
we call this new algorithm @xmath55pac ; it is defined as follows : + _ @xmath106pac : _
+ @xmath107^{m_j } pt(k-2\rightarrow k)\ ; m_{j-1}(k-2),\nonumber\end{aligned}\ ] ] where ( as before ) @xmath98 is the purification level .
@xmath108 denotes resetting the bias of spin @xmath44 to @xmath4 .
+ with three spins @xmath109^{m_1 } pt(1\rightarrow 3)m_0(1).\end{aligned}\ ] ] with five spins @xmath110^{m_2}pt(3\rightarrow 5)m_1(3).\end{aligned}\ ] ] in general , @xmath67 cools spin @xmath44 to a bias of @xmath111 , as @xmath112 is equivalent to performing the fernandez algorithm on bits @xmath113 , where bits @xmath114 and @xmath115 have equal initial biases @xmath116 .
therefore ( see details in appendix [ app : vmpac ] ) @xmath117 it is important to mention that @xmath118 could be chosen according to various criteria ; for instance , it could depend on the total number of spins @xmath50 ( or on @xmath72 ) , in addition to its dependence on @xmath98 .
the algorithm requires @xmath119 reset steps : @xmath120 hence , the order in which the @xmath118 appear is irrelevant .
[ sec : fib - variants ] the fibonacci algorithm is exhaustive , in the sense that a very large number of 3b - comp steps are performed at each recursive level .
practicable variants may be obtained by limiting the number of compressions , as we describe here .
an algorithm that reaches a distance of @xmath121 from the fibonacci series was defined as follows : + _ @xmath59-fibonacci@xcite : run @xmath122 _ @xmath123^{m_{n , k}}\ff(n , k-1),\ ] ] where @xmath124 is a reset step on bits 1 and 2 ( described by @xmath125 in section [ sec : mpac - variants ] ) , @xmath126 are chosen such that the bias of bit @xmath44 is at least @xmath127 , and @xmath128 is the @xmath129 fibonacci number . here
we choose @xmath130 , such that @xmath131 .
this condition sets @xmath132 ( see appendix [ app : fib - run ] ) . for three spins fibonacci
is @xmath133^{m_{3,3}}\ff(3,2)\\\ \ff(3,2 ) & = & m_0(1)\ ; pt(1\rightarrow 2)\ ; m_0(1).\end{aligned}\ ] ] this attributes spin number three with a bias of at least @xmath134 for four spins @xmath135^{m_{4,4}}\ff(4,3)\\\ \ff(4,3 ) & = & [ \ff(4,2)\ ; \bb(3)]^{m_{4,3}}\ff(4,2)\\\ \ff(4,2 ) & = & m_0(1)\ ; pt(1\rightarrow 2)\ ; m_0(1),\end{aligned}\ ] ] which attributes spins three and four with biases of at least @xmath136 and @xmath137 respectively .
the term @xmath59-fib hereinafter always refers to @xmath138 , such that @xmath132 . for large spin systems , say @xmath139 or so , @xmath59-fib is not practicable , as it requires many cycles in the lower recursion levels . to circumvent this problem
, we fix the number of compression steps , such that @xmath140 and denote this variant of fibonacci by mfib : _ mfib : run @xmath122 _ + @xmath141^{m}\ff(n , k-1).\ ] ] where @xmath124 is a reset step on bits 1 and 2 as before , and @xmath36 is any integer @xmath93 . with three spins @xmath142^{m}\ff(3,2),\end{aligned}\ ] ] and spin three is attributed a bias of @xmath143 with four spins @xmath144^{m}\ff(4,3)\\\ \ff(4,3 ) & = & \left[\ff(4,2)\ ; \bb(3)\right]^{m}\ff(4,2),\end{aligned}\ ] ] @xmath145 is the same as for three spins , @xmath146 and spin four is attributed a bias of : @xmath147 in general , for @xmath50 spins , the bias of the msb is given by the recursive formula : @xmath148 where @xmath149 specifically , we focus on cases where @xmath36 is a small constant ( @xmath150 ) . for @xmath151 , it can be shown that mfib is outperformed by pac3 @xcite .
the run - time of mfib is @xmath152
it is interesting to compare sopac to other algorithms .
we first consider the cooling levels attained by each algorithm . @xmath105pac attributes to spin @xmath23 a bias of @xmath153 . in comparison , with @xmath50 spins , the ppa and all - bonacci reach @xmath53 , and @xmath154 ;
numerical analysis suggests that the lower bound of @xmath53 is the asymptotic bias @xcite . ] and fibonacci asymptotically approaches @xmath155 where @xmath156 is the @xmath157 element of the fibonacci series .
while the ppa and the fibonacci algorithm cool the entire spin - system , mpac is defined so that it only polarizes the _ most significant bit ( msb)_. for a fair comparison of run - time we need to cool the entire string in mpac as well . to accomplish this , successive applications of mpac cool the less significant bits .
namely , the process @xmath158 yields the asymptotic biases @xmath159
@xmath160 @xmath161 consider the application of cooling algorithms to cool all spins ( with initial biases of 0 ) , until the biases are sufficiently close to the asymptotic biases ; the resulting biases of the first seven spins ( as long as @xmath162 of the coldest spin is still much smaller than 1 ) are given here for @xmath86pac and other exhaustive algorithms , as well as for practicable and sopac algorithms : * @xmath86pac @xmath163 * ppa and all - bonacci @xmath164 * fibonacci @xmath165 * pac2 @xmath77 1pac @xmath166 * pac3 @xmath167 * 2pac @xmath168 * 4pac @xmath169 * 6pac @xmath170 * 3fib @xmath171 * 4fib @xmath172 the bias configurations are given in units of the initial bias , @xmath4 .
@xmath59-fib is not included , as for each total number of spins it produces different bias series . for @xmath173-fib yields the biases @xmath174 .
we consider a small spin - system comprising five spins .
table [ tab : bias-5-spins ] compares the biases ( for the msb ) obtained by previous algorithms ( top ) , mpac ( middle section ) , and mfib ( bottom ) , as well as the number of resets required to create the entire bias series -bit gate is performed in a single computing step , and that the total number of such computing steps is negligible with respect to the duration of the reset steps . in reality this is not the case , and practicable algorithms such as pac and sopac are important . ] .
note that the ppa cools better when more resets are allowed , approaching the limit of @xmath175 for ac with 5 spins . with only 28 resets ,
the ppa attains a semi - optimal cooling level of @xmath176 ; for a similar cooling level , 4pac and 6pac require 101 and 197 resets , respectively .
after 99 resets , the ppa obtains a near - optimal cooling level of @xmath177 .
table [ tab : bias-7-spins ] presents a similar comparison for 7 spins .
table [ tab : bias-9-spins ] and table [ tab : bias-11-spins ] present similar comparisons for 9 and 11 spins , respectively ; a spin - system of comparable size was recently used for benchmarking quantum control methods @xcite .
appendix [ app : cool - factors ] compares the number of spins and run - time required by each algorithm to achieve several small cooling factors .
the run - time analysis of the pac algorithms is conveniently expressed as a function of the purification level , @xmath72 ( see appendix [ app : alg - pac ] ) .
the entire spin - system is cooled by @xmath178 successive applications of each algorithm , as shown above for mpac .
in contrast , the fibonacci algorithm and the ppa were designed to generate the entire series of biases .
the run - time of @xmath59-fib is given by @xmath179 ( see calculation in appendix [ app : fib - run ] ) , and the run - time of the ppa was obtained by a computer simulation that iterates between the two steps of the algorithm . . cooling a 5-spin system by various algorithms .
the optimal cooling level is @xmath175 . [
cols="<,^,^",options="header " , ] +
) for 4pac , 6pac and @xmath86pac as a function of the purification level @xmath180 starting from an initial bias of @xmath181 . for more details
see figure [ fig : mpac - analysis - for-0.01 ] . ] | _ algorithmic cooling ( ac ) of spins _ applies entropy manipulation algorithms in open spin - systems in order to cool spins far beyond shannon s entropy bound .
ac of nuclear spins was demonstrated experimentally , and may contribute to nuclear magnetic resonance ( nmr ) spectroscopy .
several cooling algorithms were suggested in recent years , including practicable algorithmic cooling ( pac ) and exhaustive ac .
_ practicable _ algorithms have simple implementations , yet their level of cooling is far from optimal ; _ exhaustive _ algorithms , on the other hand , cool much better , and some even reach ( asymptotically ) an optimal level of cooling , but they are not practicable .
we introduce here _ semi - optimal _ practicable ac ( sopac ) , wherein few cycles ( typically 26 ) are performed at each recursive level .
two classes of sopac algorithms are proposed and analyzed .
both attain cooling levels significantly better than pac , and are much more efficient than the exhaustive algorithms .
the new algorithms are shown to bridge the gap between pac and exhaustive ac .
in addition , we calculated the number of spins required by sopac in order to purify qubits for quantum computation .
as few as 12 and 7 spins are required ( in an ideal scenario ) to yield a mildly pure spin ( 60% polarized ) from initial polarizations of 1% and 10% , respectively . in the latter case ,
about five more spins are sufficient to produce a highly pure spin ( 99.99% polarized ) , which could be relevant for fault - tolerant quantum computing . |
development of the methods that allow improving accuracy of determining the asteroid sizes ( i.e. whether they measuredozens or hundreds meters in diameter ) is important for correct estimateof damage they can cause ( either regional or global catastrophes , respectively ) . at the same time
this research can be interesting for specialists who study shapes and the surface geometry of small bodies of the solar system . in our previous works
@xcite we proposed the method to estimate sizes of passive cosmic objects which method utilizes the radiolocation probing by ultra - high - resolving nanosecond signals for obtaining radar signatures .
the method involves radio pulse strobing of reflected ultra - high - resolving signals from the surface of the cosmic object .
the complete coherence of the probing and reflected signals is an essential condition of the method .
however such a condition corresponds to idealized case when no phase instabilities exist in the signal processing system .
the real sources of reference oscillations have nonzero instant instability of frequency which leads to loss of coherence at large signal lags ( large distances ) .
this factor restricts performance of coherent processing methods and leads to reduction of signal - to - noise ratio at the output of a stroboscopic system . in the analysis of time scale transformation of broadband radio
signals @xcite the complete coherence of carrier frequencies of the measured and the reference oscillations is commonly assumed .
such a concept corresponds to an absence of phase instabilities in the signal processing system . in real devices
this condition can be broken due to deviations of reference generator frequency and phase , instabilities of delays in a signal path and other factors .
these factors restrict performance of coherent processing methods and lead to reduction of signal - to - noise ratio at the output of a stroboscopic system .
let us consider the influence of phase instability of the carrier frequencies of the measured and the reference signals on statistical characteristics of the transformed signal in stroboscopic processing .
we will describe the loss of coherence by a random process @xmath0
, i.e. by a fluctuation component of a phase difference between the received and the strobe radio signals
. the statistical characteristics of the phase difference @xmath0 are considered to be known . in the analysis
we will assume that the coefficient of spectral transformation @xmath1 is large enough to use asymptotic estimates .
the model of stroboscopic processing of the reflected signals ( fig .
[ fig:1 ] ) differs from one considered in the work @xcite by the low - pass filter being replaced with the tracking filter which adaptively tunes to differential frequency of carriers @xmath2 where @xmath3 is the carrier frequency of the probing signal , @xmath4 is the radial velocity of an asteroid . for an exact determination of differential frequency @xmath5
, the radial velocity @xmath4 of an asteroid has to be measured independently using narrow - band methods based on center of mass of the doppler signal spectrum .
one of such effective methods is the method of real - time assessment of radial velocity by means of fractional differentiation of a doppler signal considered in the previous work of the authors @xcite .
let us represent the complex models of the received @xmath6 and the reference @xmath7 signals @xcite including phase instability @xmath0 in the form @xmath8 } } , \qquad \dot{a}(t ) = \sum_{k=0}^n { a_1(t - kt_1)e^{j\omega_1 t}},\ ] ] where @xmath9 and @xmath10 are the complex envelopes providing high range resolution ; @xmath11 , @xmath12 are the repetition periods of the signal and the strobe ; @xmath13 ; @xmath14 is the sampling increment ( @xmath15 ) .
the value of stroboscopic sample of a signal in the @xmath16-th sampling period can be presented as @xmath17 } dt.\ ] ] let us assume @xmath0 be a stationary random zero - mean process with correlation window @xmath18 exceeding the duration of strobe signals .
let us suppose also that slow phase displacements are tracked by stabilizing system , wherefore one can neglect the correlation of adjacent samples @xmath19 and set @xmath20 .
this allows for presentation of the sample @xmath21 in the form @xmath22}}{2 t } \int_{0}^{t } a(t^{\prime } ) a_1(t^{\prime}-k\delta t ) dt = \dot{y}_{k0 } e^{j\theta_k}.\ ] ] where @xmath23 is the sample of the random process @xmath0 and @xmath24 stands for the stroboscopic sample when there is no phase instability in the signal processing system . to ensure the mode of `` ultra - high - resolution '' of radar signatures of asteroids
@xcite one has to use nanosecond signals with pulse ratio of order @xmath25 , thus , the aforementioned approximations are perfectly acceptable . the average value ( mathematical expectation ) of samples obtained by averaging over phase @xmath26 can be expressed as @xmath27 where @xmath28 \rangle = \chi_{\theta}(1)$ ] ; @xmath29 is the characteristic function of the distribution law of phase fluctuations @xmath30 .
for the normal process with zero mean @xmath31 this value is equal to @xmath32 < 1 $ ] .
variance of samples @xmath21 amounts to @xmath33 under the assumptions made .
the ratio of variance to squared mean value is @xmath34 it has the meaning of relative power level of output noise resulting from phase fluctuations .
this ratio can be significantly reduced by increasing the spectral transformation coefficient @xmath35 at the expense of sampling step @xmath36 decrease and by using data storage in a system digital filter . in this case
the variance @xmath37 will be lowered by a factor @xmath38 where @xmath38 is the accumulation coefficient @xcite .
given that the filter and the spectrally compressed signal band are adaptively matched , the value @xmath38 is asymptotically equivalent to the number of sampling steps @xmath36 packed in the signal duration @xmath39 : @xmath40 .
( mathematical expectation and root - mean - square deviation ) vs. phase instability @xmath41 for different values of accumulation coefficient @xmath38.,scaledwidth=100.0% ]
we performed numerical simulation of processing in the radio pulse strobing scheme ( fig .
[ fig:1 ] ) of the signal @xmath42 $ ] reflected by a single bright point of the surface of a moving asteroid .
the envelopes of the probing and strobing @xmath43 signals were chosen identical : @xmath44 $ ] with the effective duration @xmath45 determined according to the method of moments @xcite : @xmath46 .
the random process @xmath0 was specified as a sequence of uncorrelated samples @xmath47 with the normal distribution : @xmath48 .
the filter s transfer function was rectangular with the bandwidth equal to the width of the transformed signal power spectrum at @xmath49 @xmath50 ( 20db ) level .
since the energy of received signal is important for optimal reception under additive noise conditions @xcite , the influence of phase instability was estimated as a decrease of the mean signal s energy @xmath51 at the output of the signal processing filter , relative to the energy @xmath52 under full coherence condition ( @xmath53 ) . fig .
[ fig:2 ] demonstrates statistical characteristics of the output signal of the stroboscopic signal processing system for different accumulation coefficients in the filter obtained by statistical modeling at specified values @xmath54 ; @xmath55 ; @xmath56 ; @xmath57 . the quantity @xmath58 corresponding to signal s energy decrease at the output of the band - pass filter of the stroboscopic signal processing system vs. phase instability is plotted in fig .
[ fig:2 ] .
the relative errors caused by phase instability are also shown .. presented results of statistical modeling are obtained by averaging over 100 simulation runs .
as previously noted @xcite , for a cosmic object of about 50 m in size the range resolution of @xmath59 m can be provided by coherent stroboscopic signal processing of signals with duration @xmath60 ns in the x - range ( @xmath61 ghz ) and frequency band @xmath62 mhz . as it can be seen from fig .
[ fig:2 ] decrease of mean received signal power by half corresponds to the phase instability of @xmath63 rad .
the value of phase deviation @xmath64 caused by a short - time frequency instability @xmath65 and by finite signal propagation time @xmath66 is @xmath67 it represents the wiener process @xcite with normal distribution .
the upper - bound estimate of phase difference gives @xmath68 .
for operation of stroboscopic radar station at range @xmath69 with acceptable phase instability of @xmath70 rad it is required to ensure @xmath71 .
this condition ensures that the noise level does not exceed @xmath72 db at the accumulation coefficient @xmath73 . at the contemporary technology level the stability of reference generators with relative error
no grater than @xmath74 , which corresponds to @xmath75 hz in the x - range , is quite realizable @xcite .
thus , the system can functionally operate within 5 million km distance .
loss of coherence in stroboscopic radar ranging systems caused by phase instabilities of the reference sources leads to sensitivity degradation and is equivalent to the effects of modulating interference .
noise reduction at the output of the stroboscopic converter caused by loss of coherence can be achieved by reducing the sample step @xmath76 with corresponding increase of the processing time .
we are grateful to vitaly korolev for the help at vectorization of drawings and victor levi for careful reading of manuscript .
the work is fulfilled within the framework of projects supported by grants from the russian foundation for basic research 15 - 47 - 02438-r - povolzhie - a and 14 - 02 - 97001-r - povolzhie - a . | we consider the problem of coherence loss in a stroboscopic high resolution radar ranging due to phase instability of the probing and reference radio signals .
requirements to the coherence of reference generators in stroboscopic signal processing system are formulated .
the results of statistical modeling are presented .
coherence loss in stroboscopic radar ranging in the problem of asteroid size estimation v. d. zakharchenko , i. g. kovalenko , v. yu .
ryzhkov _ volgograd state university , universitetskij pr .
, 100 , volgograd 400062 , russia _ _ keywords _ near - earth objects , asteroids , stroboscopic radar observations , wideband radio signals |
the @xmath0 model , in @xmath1 dimensions and at nonzero temperature @xmath3 , is defined by the following generating functional : @xmath4 \text { , } \label{z}%\ ] ] whereas @xmath5 is the dimensionless coupling constant , @xmath6 is a normalization constant , and @xmath7 is a simple free ( euclidean ) lagrangian @xmath8 , with @xmath9 .
@xmath10 consists of @xmath11 scalar real fields , which for future convenience we denote as @xmath12 .
the fields are constrained by the condition @xmath13 which is incorporated by the delta function in eq .
( [ z ] ) .
this nonlinear constraint enforces the thermodynamics of the model on an @xmath14 dimensional hypersphere and induces the interactions between the fields .
this model is interesting because it shares some properties with yang - mills theories in 4 dimensions , see the review paper @xcite and refs .
therein : \(i ) the coupling constant @xmath5 is dimensionless and the the theory is renormalizable .
it turns out that it has a negative @xmath15 function , thus showing asymptotic freedom .
\(ii ) the model is conformal invariant at the classical level , but , just as yang - mills theories in four dimensions , at the quantum level an energy scale emerges due to quantum corrections and a non - perturbative mass - gap emerges ( trace anomaly ) @xcite : although the lagrangian @xmath7 in eq .
( [ z ] ) describes @xmath11 massless fields , a nonzero mass @xmath16 is generated dynamically due to the interactions , where @xmath17 is the renormalization parameter . since the mass is non - analytic in @xmath5 , it would vanish in perturbation theory about @xmath18 .
\(iii ) for @xmath19 the model has instantons and , at finite temperature , calorons @xcite .
it is then possible to study in a simplified framework what is the role of these nonperturbative field configurations in thermodynamics ( for the related case in ym theories see for instance refs .
@xcite and refs .
therein ) .
the aim of this work is to present an analytical study of the nonlinear @xmath0 model in @xmath1 dimensions at nonzero temperature .
we compute the pressure by employing the cjt formalism @xcite and using the auxiliary field method in the one - loop approximation ( for other thermodynamical quantities and the two - loop calculation and technical details see ref .
@xcite ) . for the case
@xmath19 we compare the pressure with the results of a first - principles monte carlo lattice calculation using the integral method . at the one - loop level
, a good agreement is found for small temperature , but the analytic result for the pressure is too large when @xmath3 increases . when going to the two - loop approximation , a considerable improvement in the high @xmath3 domain is obtained , although the low-@xmath3 part is slightly worsened .
following ref . @xcite we rewrite the partition function in eq .
( [ z ] ) by using the mathematically well - defined ( i.e. convergent ) form of the @xmath20-function@xmath21 \right\ } } \text { } % \ ] ] thus obtaining @xmath22}$ ] , where@xmath23 the quantity @xmath24 is an auxiliary field serving as a lagrange multiplier . upon shifting the fields @xmath25 and @xmath24 by their condensates , @xmath26 and @xmath27 ,
a bilinear mixing term of the type @xmath28 arises .
it can be eliminated by a further shift of the @xmath24 field , @xmath29 in this way non - diagonal propagators are avoided .
\(i ) @xmath49 @xmath50 : @xmath51 ^{-1}\text { .}\nonumber\ ] ] \(ii ) @xmath52 @xmath53 : @xmath54 ^{-1}\text { .}\nonumber\ ] ] there is no solution to the gap equations of case ( i ) , since the integral @xmath55^{-1}}{2\pi k } d k\ ] ] is divergent , whereas @xmath56 is finite .
therefore , contrary to the four - dimensional case @xcite , case ( ii ) is the right choice .
this means that in two dimensions there is no spontaneous symmetry breaking of the global @xmath0 symmetry of the nonlinear @xmath0 model .
this reflects the mermin - wagner - coleman theorem , see ref .
@xcite , which forbids spontaneous breakdown of a continuous symmetry in a homogeneous system in one spatial dimension .
the model needs to be regularized and then renormalized .
here we just report the results of these steps and we refer to ref .
@xcite for a detailed treatment .
the renormalized coupling constant @xmath57 develops a dependence on the renormalization scale @xmath17 : @xmath58 thus showing that the theory is asymptotically free . at nonzero
@xmath3 the mass of the excitations is @xmath59 for @xmath60 we obtain @xmath61 which shows that a mass gap is generated and that the dilatation symmetry is broken by quantum fluctuations ( the energy scale @xmath62 is dynamically generated )
. then the function @xmath63 rises linearly with @xmath3 when @xmath3 increases : this is an expected properties for a gas of quasiparticles and takes place also in the deconfined phase of yang - mills theories , e.g. ref .
@xcite and refs . therein . .
the upper right line marks the stefan - boltzmann limit @xmath64.,scaledwidth=66.0% ] we now turn to the pressure of the system , which is , up to a sign , identical to the minimum of the effective potential : @xmath65 .
its renormalized form reads@xmath66 once the pressure is known one can compute the energy density @xmath67 and the trace anomaly @xmath68 using the first principle of thermodynamics : @xmath69 , @xmath70 . the one - loop pressure of eq .
( [ pressure ] ) is reported in fig .
[ fig : pressure ] and is compared with a lattice study of this system @xcite for the case @xmath71 one can notice that the result is good for small temperatures , but it does not match the lattice data when the temperature increases .
this result is indeed expected : the one - loop expression in eq .
( [ pressure ] ) reduces to @xmath72 for large @xmath3 , which corresponds to a free gas of @xmath11 non - interacting particles .
this result can not be correct , because the constraint expressed by the @xmath20 function in eq .
( [ z ] ) eliminates one degree of freedom .
one can go beyond the one - loop results and study the pressure at the two - loop level using the formalism developed in @xcite , see also ref .
a better agreement is obtained for large @xmath73 although the low-@xmath3 domain is slightly worsened .
indeed , it is possible to show analytically that for large @xmath3 the two - loop pressure approaches the correct value @xmath74 .
in this work we have studied the @xmath0 model in 1 + 1 dimensions at nonzero @xmath75 we have described how a mass gap is generated for @xmath60 , thus demonstrating the occurrence of dimensional transmutation .
we have then shown explicitly the analytic results for the pressure at the one - loop level and briefly commented on the results at the two - loop level .
finally , these analytic expressions for the pressure have been compared with lattice results , see fig .
[ fig : pressure ] .
more advanced analytical calculations can be performed in the future and compared to lattice data : in fact , this system represents a good tool to test nonperturbative approaches , which can then be applied to the more difficult case of yang - mills theories in four dimensions .
v. a. novikov , m. a. shifman , a. i. vainshtein and v. i. zakharov , phys .
rept . * 116 * ( 1984 ) 103 [ sov .
j. part .
* 17 * ( 1986 ) 204 ] ; fiz .
chast . atom .
yadra * 17 * ( 1986 ) 472 .
u. wolff , nucl .
b * 334 * ( 1990 ) 581 .
f. bruckmann , phys .
* 100 * ( 2008 ) 051602 [ arxiv:0707.0775 [ hep - th ] ] ; j. o. andersen , d. boer and h. j. warringa , phys
. rev .
d * 74 * ( 2006 ) 045028 [ hep - th/0602082 ] ; i. affleck , phys .
b * 92 * ( 1980 ) 149 .
r. hofmann , int .
j. mod .
phys . a * 20 * ( 2005 ) 4123 [ hep - th/0504064 ]
; u. herbst and r. hofmann , isrn high energy phys .
* 2012 * ( 2012 ) 373121 [ hep - th/0411214 ] ; f. giacosa and r. hofmann , prog .
* 118 * ( 2007 ) 759 [ hep - th/0609172 ] .
f. giacosa , phys .
d * 83 * ( 2011 ) 114002 [ arxiv:1009.4588 [ hep - ph ] ] ; f. brau and f. buisseret , phys . rev .
d * 79 * ( 2009 ) 114007 [ arxiv:0902.4836 [ hep - ph ] ] ; a. peshier , b. kampfer , o. p. pavlenko and g. soff , phys . rev .
d * 54 * ( 1996 ) 2399 ; p. castorina and m. mannarelli , phys .
c * 75 * ( 2007 ) 054901 [ hep - ph/0701206 [ hep - ph ] ] . | the @xmath0 model in @xmath1 dimensions presents some features in common with yang - mills theories : asymptotic freedom , trace anomaly , non - petrurbative generation of a mass gap .
an analytical approach to determine the termodynamical properties of the @xmath2 model is presented and compared to lattice results . here
the focus is on the pressure : it is shown how to derive the pressure in the cjt formalism at the one - loop level by making use of the auxiliary field method .
then , the pressure is compared to lattice results . address= institut fr theoretische physik , johann wolfgang goethe - universitt , + max - von - laue - str . 1 , d-60438 frankfurt , germany address= institut fr theoretische physik , johann wolfgang goethe - universitt , + max - von - laue - str . 1
, d-60438 frankfurt , germany address= institut fr theoretische physik , johann wolfgang goethe - universitt , + max - von - laue - str . 1 , d-60438 frankfurt , germany address= institut fr theoretische physik , johann wolfgang goethe - universitt , + max - von - laue - str . 1 , d-60438 frankfurt , germany , altaddress= fakultt fr physik , universitt bielefeld , + universittsstr .
25 , d-33615 bielefeld , germany |
the description of transport properties for states close to equilibrium in gaseous binary mixtures is well established . in these situations ,
the curie principle @xcite states that the presence of a velocity gradient ( second - rank tensorial quantity ) can not modify a vectorial quantity such as the mass flux @xmath0 , which is generated by gradients of mole fraction @xmath1 , pressure @xmath2 , and temperature @xmath3 . as a consequence , the mutual diffusion coefficient @xmath4 ( which couples the mass current with @xmath5 ) , the pressure diffusion coefficient @xmath6 ( which couples the mass current with @xmath7 ) and the thermal diffusion coefficient @xmath8 ( which couples the mass current with @xmath9 ) do not depend on the velocity gradient .
however , when the shear rate applied is large , non - newtonian effects are important so that the curie principle does not hold and the coefficients associated with the mass transport are affected by the presence of shear flow . in particular , if the spatial gradients @xmath5 , @xmath7 , and @xmath9 are weak , one expects that the flux @xmath0 is still linear in these gradients but the standard scalar coefficients @xmath10 must be replaced by the shear - rate dependent second - rank tensors @xmath11 . the aim of this paper is to determine the above tensors in the framework of the boltzmann equation .
we are interested in a situation where _ weak _ spatial gradients of mole fraction , pressure , and temperature coexist with a _ strong _ shear rate . under these conditions ,
the application of the conventional chapman - enskog expansion @xcite around the local equilibrium state to get higher order hydrodynamic effects ( burnett , super - burnett , @xmath12 ) to the mass flux turns out to be extremely difficult .
this gives rise to look for alternative approaches .
a possibility is to expand around a more relevant reference state than local equilibrium .
since we want to compute the mass transport in a strongly sheared mixture , the so - called uniform shear flow ( usf ) state can be chosen as the reference state .
the usf state is characterized by constant mole fractions , a uniform temperature , and a linear velocity profile @xmath13 , where @xmath14 is the constant shear rate .
due to its simplicity , this state has been widely used in the past to shed light on the complexities associated with the _ nonlinear _ response of the system to the action of strong shearing .
in addition , the usf state is one of the rare exceptions for which the hierarchy of moments of the boltzmann equation admits an exact solution for single @xcite and multicomponent gases @xcite of maxwell molecules ( repulsive potential of the form @xmath15 ) . in this case ,
explicit expressions of the pressure tensor ( which is the relevant irreversible flux of the problem ) have been obtained for _ arbitrary _ values of the shear rate and the parameters of the system ( masses , concentrations and force constants ) .
as said before , here we want to compute the mass transport under usf for maxwell molecules . since the mixture is slightly perturbed from the usf , the boltzmann equation can be solved by an expansion in small gradients around the ( local ) shear flow distribution instead of the ( local ) equilibrium .
this is the main feature of the expansion since the reference state is not restricted to small values of the shear rate . in the first order of the expansion ,
the set of generalized transport coefficients @xmath11 are identified from the mass flux @xmath0 as nonlinear functions of the shear rate and the parameters of the mixture .
this chapman - enskog - like expansion has been used to analyze transport properties in spatially inhomogeneous states near usf in the case of ordinary gases @xcite and more recently in the context of granular gases .
@xcite some previous attempts have been carried out earlier by the author and coworkers @xcite in the case of the diffusion tensor @xmath16 . however , all these studies have been restricted to perturbed _ steady _ states with the constraints @xmath17 and @xmath18 .
although steady states are in general desirable for practical purposes , especially in computer simulations , @xcite here we extend the above studies to a general time and space dependence of the hydrodynamic fields .
this allows us to evaluate new contributions to the mass flux ( those proportional to @xmath7 and @xmath9 ) , which where not taken into account in the previous studies .
@xcite the plan of the paper is as follows .
first , a brief summary of the results obtained from the boltzmann equation for a binary mixture of maxwell molecules under usf is presented in sec .
section [ sec3 ] deals with the perturbation scheme used to solve the boltzmann equation for the mixture to first order in the deviations of the hydrodynamic field gradients from their values in the reference shear flow state .
the generalized transport coefficients characterizing the mass transport around usf are also defined in sec.[sec3 ] .
these coefficients are explicitly obtained in sec .
[ sec4 ] with and without the presence of an external thermostat introduced usually in computer simulations to compensate for the viscous heating .
the dependence of some of these coefficients on the shear rate is illustrated with detail in the tracer limit case , showing that the influence of shear flow on mass transport is quite significant .
the paper is closed by a brief discussion of the results in sec .
[ sec5 ] , the generalized transport coefficients associated with the momentum and heat transport are evaluated in appendix [ appd ] from a simple model kinetic equation of the boltzmann equation .
we consider a _ dilute _ binary mixture where @xmath19 is the one - particle velocity distribution function of species @xmath20 ( @xmath21 ) .
the time evolution of the distributions @xmath22 is given by the set of two coupled nonlinear boltzmann equations : @xmath23 \ ; , \label{2.1}\ ] ] where @xmath24 is the mass of a particle of species @xmath20 , @xmath25 is a possible external force acting on particles of species @xmath20 , and @xmath26 $ ] is the boltzmann collision operator , which in standard notation reads @xcite @xmath27=\int d{\bf v}_1 \int d\omega |{\bf v}-{\bf v}_1|\sigma_{sr}({\bf v}-{\bf v}_1,\theta ) [ f_s({\bf v } ' ) f_r({\bf v}'_1)-f_s({\bf v } ) f_r({\bf v}_1 ) ] \;.\ ] ] the basic moments of @xmath22 are the species number densities @xmath28 and the mean velocity of species @xmath20 @xmath29 these quantities define the total number density @xmath30 and the flow velocity @xmath31 , where @xmath32 is the mass density of species @xmath20 and @xmath33 is the total mass density . the temperature @xmath3 is defined as @xmath34 where @xmath35 is the boltzmann constant and @xmath36 is the peculiar velocity .
the second identity in ( [ 2.5 ] ) defines the partial kinetic temperatures @xmath37 of species @xmath20 .
they measure the mean kinetic energy of particles of species @xmath20 .
moreover , in a dilute gas the hydrostatic pressure @xmath2 is given by @xmath38 .
the quantities @xmath39 , @xmath40 , and @xmath3 are associated with the densities of conserved quantities ( mass of each species , total momentum , and total energy ) .
the corresponding balance equations define the dissipative fluxes of mass @xmath41 momentum ( pressure tensor ) , @xmath42 and energy ( heat flux ) @xmath43 the second equalities in eqs .
( [ 2.7 ] ) and ( [ 2.8 ] ) define the partial contributions @xmath44 and @xmath45 to the pressure tensor and heat flux , respectively .
the fact that the mass flux @xmath46 is defined with respect to the local center - of - mass velocity @xmath40 implies that @xmath47 the usf state is macroscopically defined by constant densities @xmath39 , a spatially uniform temperature @xmath48 and a linear velocity profile @xmath49 , where @xmath14 is the _ constant _ shear rate .
since @xmath39 and @xmath3 are uniform , then @xmath50 , and the transport of momentum ( measured by the pressure tensor ) is the relevant phenomenon . in the usf problem , the temperature tends to increase in time due to viscous heating .
usually , an external force ( thermostat ) is introduced in computer simulations to remove this heating effect and keep the temperature constant .
@xcite the simplest choice is a gaussian isokinetic thermostat given by @xmath51 where the thermostat parameter @xmath52 is a function of the shear rate adjusted as to keep the temperature constant .
the implicit assumption behind the introduction of these forces is that they play a neutral role in the transport properties , so that the latter are the same with and without a thermostat , when conveniently scaled with the thermal speed . nevertheless , this expectation is not in general true , except for some specific situations and/or interaction potentials .
@xcite at a microscopic level , the usf is characterized by a velocity distribution function that becomes _ uniform _ in the local lagrangian frame , i.e. , @xmath53 . in that case , eq .
( [ 2.1 ] ) with the choice ( [ 2.9 ] ) reduces to @xmath54+j_{12}[f_1,f_2]\ ] ] and a similar equation for @xmath55 . here , @xmath56 .
the hierarchy of velocity moments associated with the boltzmann equation ( [ 2.10 ] ) can be recursively solved in the particular case of maxwell molecules , i.e. , when particles of species @xmath57 and @xmath20 interact through a potential of the form @xmath58 .
the key point is that for this interaction the collision rate @xmath59 is independent of the relative velocity @xmath60 and so the collisional moments of order @xmath61 only involve moments of degree smaller than or equal to @xmath61 . in particular ,
the first- and second - degree collisional moments are given by @xcite @xmath62=-\frac{\lambda_{sr}}{m_sm_r}\left(\rho_s{\bf j}_r-\rho_r{\bf j}_s\right),\ ] ] @xmath63&=&\frac{\lambda_{sr}'}{(m_s+m_r)m_s}\left[\left(\rho_sp_r+\rho_rp_s-\frac{2}{3}{\bf j}_s\cdot { \bf j}_r\right)\openone \right.\nonumber\\ & & \left.-\rho_s{\sf p}_r-\rho_r{\sf p}_s+{\bf j}_s{\bf j}_r+{\bf j}_r{\bf j}_s\right]\nonumber\\ & & -\frac{\lambda_{sr}}{(m_s+m_r)m_s}\left[2\left(\frac{m_s}{m_r}\rho_r{\sf p}_s-\rho_s{\sf p}_r\right)\right . \nonumber\\ & & \left.+\left(1-\frac{m_s}{m_r}\right)({\bf j}_s{\bf
j}_r+{\bf j}_r{\bf j}_s)\right],\end{aligned}\ ] ] where @xmath64 is the partial hydrostatic pressure and @xmath65 thanks to the above property , exact expressions of the pressure tensor @xmath66 for a binary mixture of maxwell molecules under usf were obtained some time ago .
@xcite the nonzero elements of @xmath66 are related to the rheological properties of the mixture , namely , the nonlinear shear viscosity and the viscometric functions . in reduced units , they turn out to be nonlinear functions of the ( reduced ) shear rate @xmath67 ( where @xmath68 is a convenient time unit defined below ) and the parameters of the mixture : the mass ratio @xmath69 , the mole fraction @xmath70 and the force constant ratios @xmath71 and @xmath72 .
it must be noted that in the particular case of maxwell molecules there is an exact equivalence between the usf results with and without the external forces ( [ 2.9 ] ) .
as will be shown below , beyond the usf problem , the presence of the thermostat does not play a neutral role in the results and a certain influence may exist .
as said in the introduction , the main aim of this work is to analyze mass transport of a dilute binary mixture subjected to usf . in that case , let us assume that the usf state is disturbed by small spatial perturbations .
the response of the system to those perturbations gives rise to contributions to the mass flux that can be characterized by generalized transport coefficients .
this section is devoted to the evaluation of those coefficients . in order to analyze this problem
we have to start from the set of boltzmann equations ( [ 2.1 ] ) with a general time and space dependence .
let @xmath73 be the flow velocity of the _ undisturbed _ usf state , where the elements of the tensor @xmath74 are @xmath75 .
in the _ disturbed _ state , however the true velocity @xmath40 is in general different from @xmath76 , i.e. , @xmath77 , @xmath78 being a small perturbation to @xmath76 . as a consequence ,
the true peculiar velocity is now @xmath79 , where @xmath80 . in the lagrangian frame moving with @xmath76 , the boltzmann equations ( [ 2.1 ] ) can be written as @xmath81+j_{12}[f_1,f_2],\ ] ] @xmath82+j_{21}[f_2,f_1],\ ] ] where here the derivative @xmath83 is taken at constant @xmath84 .
in addition , in eqs .
( [ 3.1 ] ) and ( [ 3.2 ] ) the thermostat force has been assumed to be proportional to the actual peculiar velocity , @xmath85 where now the parameter @xmath52 is in general a function of @xmath86 and @xmath87 through their functional dependence on the hydrodynamic fields @xmath39 and @xmath3 . the generalization of @xmath52 to the inhomogeneous case is essentially a matter of choice . here , for the sake of simplicity
, we will take two different choices for @xmath52 : ( i ) @xmath88 , so that the temperature grows in time , and ( ii ) the same expression obtained in the ( pure ) usf problem , except that the densities and temperature are replaced by those of the general inhomogeneous state . the macroscopic balance equations associated with this disturbed usf state are obtained by taking moments in eqs . ( [ 3.1 ] ) and ( [ 3.2 ] ) with the result @xmath89 @xmath90 @xmath91 where the mass flux @xmath46 , the pressure tensor @xmath66 , and the heat flux @xmath92 are defined by eqs.([2.6 ] ) , ( [ 2.7 ] ) , and ( [ 2.8 ] ) , respectively , with the replacement @xmath93 .
the corresponding balance equations for the mole fraction @xmath70 and the pressure @xmath38 can be obtained from eqs .
( [ 3.3 ] ) and ( [ 3.5 ] ) .
they are given by @xmath94 @xmath95 we assume that the deviations from the usf state are small .
this means that the spatial gradients of the hydrodynamic fields are small . for systems near equilibrium ,
the specific set of gradients contributing to each flux is restricted by fluid symmetry , onsager relations , and the form of entropy production .
@xcite however , in far from equilibrium situations ( such as the one considered in this paper ) , only fluid symmetry applies and so there is more flexibility in the representation of the heat and mass fluxes since they can be defined in a variety of equivalent ways depending on the choice of hydrodynamic gradients used .
in fact , some care is required in comparing transport coefficients in different representations using different independent gradients for the driving forces . here , as in previous works , @xcite the mole fraction @xmath1 , the pressure @xmath2 , the temperature @xmath3 , and the local flow velocity @xmath78
are chosen as hydrodynamic fields .
since the system is strongly sheared , a solution to the set of boltzmann equations ( [ 3.1 ] ) and ( [ 3.2 ] ) can be obtained by means of a generalization of the conventional chapman - enskog method @xcite in which the velocity distribution function is expanded around a _ local _ shear flow reference state in terms of the small spatial gradients of the hydrodynamic fields relative to those of usf .
this is the main new ingredient of the expansion .
this type of chapman - enskog - like expansion has been already considered to get the set of shear - rate dependent transport coefficients @xcite in thermostatted shear flow problems and it has also been recently used for inelastic gases .
@xcite in the context of the chapman
enskog method , @xcite we look for a _
normal _ solution of the form @xmath96,\ ] ] where @xmath97 this special solution expresses the fact that the space dependence of the reference shear flow is completely absorbed in the relative velocity @xmath84 and all other space and time dependence occurs entirely through a _ functional _ dependence on the fields @xmath98 .
the functional dependence ( [ 3.5 ] ) can be made local by an expansion of the distribution function in powers of the hydrodynamic gradients : @xmath99 = f_s^{(0)}({\bf v})+ f_s^{(1)}({\bf v})+\cdots,\ ] ] where the reference zeroth - order distribution function corresponds to the usf distribution function but taking into account the local dependence of the concentration , pressure and temperature and the change @xmath100 .
the successive approximations @xmath101 are of order @xmath61 in the gradients of @xmath1 , @xmath2 , @xmath3 , and @xmath78 but retain all the orders in the shear rate @xmath14 . here , only the first - order approximation will be analyzed .
when the expansion ( [ 3.8 ] ) is substituted into the definitions ( [ 2.6 ] ) , ( [ 2.7 ] ) , and ( [ 2.8 ] ) , one gets the corresponding expansions for the fluxes : @xmath102 @xmath103 finally , as in the usual chapman - enskog method , the time derivative is also expanded as @xmath104 where the action of each operator @xmath105 is obtained from the hydrodynamic equations ( [ 3.3])([3.5 ] ) .
these results provide the basis for generating the chapman - enskog solution to the boltzmann equations ( [ 3.1 ] ) and ( [ 3.2 ] ) . substituting the expansions ( [ 3.9])([3.11 ] ) into eq .
( [ 3.1 ] ) , the kinetic equation for @xmath106 is given by @xmath107+j_{12}[f_1^{(0)},f_2^{(0)}].\ ] ] to lowest order in the expansion the conservation laws give @xmath108 @xmath109 if @xmath88 , then @xmath110 while if @xmath111 then @xmath112 . since @xmath106 is a normal solution , the time derivative in eq .
( [ 3.12 ] ) can be represented more usefully as @xmath113 where in the last step we have taken into account that @xmath106 depends on @xmath78 only through the peculiar velocity @xmath114 .
substituting eq .
( [ 3.15 ] ) into eq . ( [ 3.12 ] ) yields the following kinetic equation for @xmath106 : @xmath115+j_{12}[f_1^{(0)},f_2^{(0)}].\end{aligned}\ ] ] a similar equation holds for @xmath116 .
the partial pressure tensors @xmath117 and @xmath118 can be obtained from eq .
( [ 3.16 ] ) and its counterpart for @xmath116 when one multiplies both equations by @xmath119 and integrate over @xmath114 .
their explicit forms can be found in the appendix of ref .. the analysis to first order in the gradients is worked out in appendix [ appa ] .
the distribution function @xmath120 is of the form @xmath121 where the vectors @xmath122 , and the tensor @xmath123 are functions of the true peculiar velocity @xmath114 .
they are the solutions of the following set of linear integral equations : @xmath124 @xmath125{\boldsymbol { \cal b}}_{1}={\bf b}_{1}-\left[\frac{2at}{3p^2}(1-p\partial_{p})p_{xy}^{(0)}-2t\partial_{p}\alpha \right]{\boldsymbol { \cal c}}_1,\end{aligned}\ ] ] @xmath126{\boldsymbol { \cal c}}_{1}={\bf c}_{1}+\left(\frac{2a}{3}\partial_{t}p_{xy}^{(0)}+2p\partial_{t}\alpha \right){\boldsymbol { \cal b}}_1,\end{aligned}\ ] ] @xmath127 where @xmath128 , @xmath129 , @xmath130 , and @xmath131 are defined by eqs .
( [ a8])([a11 ] ) , respectively .
in addition , @xmath132 and @xmath133 are the linearized boltzmann collision operators around the reference usf state : @xmath134+j_{11}[x , f_1^{(0)}]+j_{12}[x , f_2^{(0)}]\right),\ ] ] @xmath135.\ ] ] in this paper we are mainly interested in evaluating the first - order contribution to the mass flux @xmath136 .
it is defined as @xmath137 use of eq .
( [ 3.17 ] ) into eq .
( [ 3.24 ] ) gives the expression @xmath138 where @xmath139 @xmath140 @xmath141 upon writing eqs .
( [ 3.26])([3.28 ] ) use has been made of the symmetry properties of @xmath142 , @xmath143 , and @xmath144 . in general , the set of _ generalized _ transport coefficients @xmath16 , @xmath145 , and @xmath146 are nonlinear functions of the shear rate and the parameters of the mixture .
it is apparent that the anisotropy induced by the presence of shear flow gives rise to new transport coefficients for the mass flux , reflecting broken symmetry .
according to eq.([3.25 ] ) , the mass flux is expressed in terms of a diffusion tensor @xmath16 , a pressure diffusion tensor @xmath145 , and a thermal diffusion tensor @xmath146 . to get the explicit dependence of the above transport coefficients on the parameter space of the problem
, the form of @xmath52 must be chosen .
as said before , two choices will be considered here : ( i ) the unthermostatted case @xmath88 , and ( ii ) the thermostatted case @xmath111 .
both cases will be separately studied in the next section .
this section is devoted to the determination of the generalized transport coefficients @xmath16 , @xmath145 , and @xmath146 associated with the mass transport for the two choices of the thermostat parameter .
these coefficients are given in terms of the solutions to the integral equations ( [ 3.18])([3.20 ] ) . in the absence of an external thermostat ( @xmath88 ) , the integral equations ( [ 3.18])([3.20 ] ) become @xmath147 @xmath148 @xmath149{\boldsymbol { \cal c}}_{1}+{\cal l}_1{\boldsymbol { \cal c}}_{1}+{\cal m}_1{\boldsymbol { \cal c}}_{2}\nonumber\\ & & = { \bf c}_{1}+\frac{2a}{3}{\boldsymbol { \cal b}}_1(\partial_{t}p_{xy}^{(0)}).\end{aligned}\ ] ] the dependence of @xmath150 on the pressure @xmath2 and temperature @xmath3 occurs explicitly and through its dependence on the reduced shear rate @xmath67 . here
, the effective collision frequency @xmath68 is given by @xcite @xmath151 where @xmath152 is defined in eq .
( [ 4.11 ] ) .
consequently , @xmath153 @xmath154 where @xmath155 .
in addition , the dependence of @xmath150 on the mole fraction @xmath1 is also rather intricate and so the derivatives with respect to @xmath1 must be carried out with care . @xcite the generalized coefficients @xmath16 , @xmath145 , and @xmath146 can be obtained from eqs.([4.1])([4.3 ] ) when one multiplies those equations by @xmath156 and integrates over @xmath114 .
after some algebra ( some technical details are provided in appendix [ appc ] ) , one arrives at the following set of coupled algebraic equations : @xmath157d_{kj}&=&\frac{\rho
k_bt}{m_1m_2}\left(\partial_{x_1}p_{1,ij}^{*}- \frac{\rho_1}{\rho}\partial_{x_1}p_{ij}^*\right)\nonumber\\ & & + \frac{2a\rho^2}{3m_1m_2n}(\partial_{x_1}p_{xy}^ { * } ) \left(d_{p , ij}+d_{t , ij}\right),\end{aligned}\ ] ] @xmath158d_{p , kj}&= & \frac{p}{\rho}(1-a^*\partial_{a^ * } ) \left(p_{1,ij}^*-\frac{\rho_1}{\rho}p_{ij}^*\right)\nonumber\\ & & -\frac{2a}{3}a^*d_{t , ij}(\partial_{a^*}p_{xy}^*),\end{aligned}\ ] ] @xmath159d_{t , kj}&= & \frac{p}{\rho}a^*\partial_{a^ * } \left(p_{1,ij}^*-\frac{\rho_1}{\rho}p_{ij}^*\right)\nonumber\\ & & + \frac{2a}{3}a^*d_{p , ij}(\partial_{a^*}p_{xy}^*),\end{aligned}\ ] ] where @xmath160 and use has been made of the relations ( [ 4.5 ] ) and ( [ 4.6 ] ) . in the absence of shear field ( @xmath161 ) , then @xmath162 , and @xmath163 , so that eqs . (
[ 4.23])([4.25 ] ) have the solutions @xmath164 , @xmath165 , and @xmath166 , where @xmath167 and @xmath168 are the conventional navier - stokes transport coefficients for maxwell molecules .
@xcite their expressions are @xmath169 in this case the mass flux @xmath170 can be written as @xcite @xmath171 where @xmath172 @xmath173 being the chemical potential per unit mass .
the fact that the thermal diffusion coefficient vanishes when @xmath174 is due to the interaction potential considered ( maxwell molecules ) since this coefficient is different from zero for more general interaction potentials .
@xcite however , when the mixture is strongly sheared , the boltzmann equation leads to contributions to the mass flux proportional to the thermal gradient , even for maxwell molecules . in the case of mechanically equivalent particles ( @xmath175 , @xmath176 ) , @xmath177 , @xmath178 , and so @xmath179 .
moreover , eq . ( [ 4.23 ] )
reduces to @xmath180 equation ( [ 4.26.3 ] ) is consistent with previous results derived for the self - diffusion tensor .
@xcite furthermore , known results for the diffusion tensor @xcite are also recovered in the tracer limit ( @xmath181 ) .
let us assume now that an external thermostat is introduced to compensate for the viscous heating effect
. in this case , @xmath111 , and the integral equations ( [ 3.18])([3.20 ] ) become @xmath182 @xmath183 @xmath184 in contrast to what happens in the unthermostatted case , the different integral equations are now decoupled and hence the generalized coefficients of the mass transport can be obtained more easily . the mathematical steps to get them are similar to those made before when @xmath88 and so only the final results are presented .
the explicit expressions for @xmath16 , @xmath145 , and @xmath146 are given by @xmath185 @xmath186 @xmath187 in order to get these expressions use has been made of the identity @xmath188 where @xmath189 is an arbitrary constant and @xmath74 is the tensor with elements @xmath75 . in the case of mechanically equivalent particles , @xmath179 and eq .
( [ 4.30 ] ) reduces to @xmath190 equation ( [ 4.33 ] ) gives the self - diffusion tensor of tagged particles under thermostatted usf .
@xcite for a general binary mixture , the expression ( [ 4.30 ] ) for the diffusion tensor @xmath16 coincides with the one derived before @xcite in a stationary state with the constraints @xmath17 and @xmath18 . finally , it is also apparent that , except for vanishing shear rates , the expressions of the generalized transport coefficients ( [ 4.30])([4.32 ] ) in the thermostatted state differ from the ones derived in the absence of a thermostat , eqs .
( [ 4.23])([4.25 ] ) .
this shows again that the presence of the thermostat affects the transport properties of the system .
the results obtained in the preceding section give all the relevant information on the influence of shear flow on the mass transport .
in general , the elements @xmath16 , @xmath145 and @xmath146 present a complex dependence on the shear rate and the parameters of the mixture without any restriction on their values .
however , although the solution to eqs .
( [ 4.23])([4.25 ] ) ( in the unthermostatted case ) and eqs.([4.30])([4.32 ] ) ( in the thermostatted case ) is simple , it involves quite a tedious algebra due to the complex dependence of the partial pressure tensors @xmath191 and the thermostat parameter @xmath52 on the mole fraction @xmath1 and the reduced shear rate @xmath192 . to show the shear - rate dependence of the tensors @xmath193 in a clearer way , the tracer limit ( @xmath181 ) will be considered here in detail . in addition , to make some contact with computer simulation results , the thermostatted case will be studied . in the tracer limit case ,
@xmath194 and the partial pressure tensors @xmath195 and @xmath196 have a more simplified forms .
@xcite in particular , @xmath197 and @xmath198 .
the explicit expressions of the partial pressure tensors in the tracer limit are provided in appendix [ appb ] . as expected , @xmath199 , in agreement with the symmetry of the problem .
as a consequence , there are five relevant elements : the three diagonal ( @xmath200 , @xmath201 , and @xmath202 ) and two off - diagonal elements ( @xmath203 and @xmath204 ) . in addition , @xmath205 and @xmath206 .
the equality @xmath207 implies @xmath208 .
this property is a consequence of the interaction model considered since for non - maxwell molecules computer simulations show that the @xmath209 and @xmath210 elements of the pressure tensor are different . @xcite in figs .
[ fig1][fig6 ] , the relevant elements of tensors @xmath211 , @xmath212 and @xmath213 are plotted as functions of the reduced shear rate @xmath192 for @xmath214 and several values of the mass ratio @xmath215 . here , the tensors have been reduced with respect to their navier - stokes values ( except @xmath146 ) , namely , @xmath216 , @xmath217 and @xmath218 .
one third of the trace of these tensors is plotted in figs .
[ fig1][fig3 ] , while the @xmath219 element is plotted in figs .
[ fig4][fig6 ] .
we observe that in general the influence of shear flow on the mass transport is quite important .
it is also apparent that the anisotropy of the system , as measured by the traces @xmath220 , @xmath221 , and @xmath222 , grows with the shear rate .
this anisotropy is more significant when the impurity is heavier than the particles of the gas .
moreover , the shear field induces cross effects in the diffusion of particles .
this is measured by the ( reduced ) off - diagonal elements @xmath223 , @xmath224 and @xmath225 . these coefficients give the transport of mass along the @xmath226 axis due to gradients parallel to the @xmath227 axis . while @xmath223 and @xmath224 are negative , the coefficient @xmath228 can be positive in the region of small shear rates .
we observe that , regardless of the mass ratio , the shapes of @xmath223 and @xmath224 are quite similar : there is a region of values of @xmath192 for which @xmath229 and @xmath230 increase with increasing shear rate , while the opposite happens for larger shear rates .
the magnitude of @xmath225 is smaller than that of the elements @xmath229 and @xmath230 , especially when the tracer particles are lighter than the particles of the gas . in this latter case
, @xmath225 is practically negligible .
diffusion of particles in a binary mixture in non - newtonian regimes is a subject of great interest from a fundamental and practical points of view .
if the mixture is strongly sheared , the mass flux @xmath0 can be significantly affected by the presence of shear flow so that the corresponding transport coefficients may differ significantly from their equilibrium values .
in addition , the resulting mass transport is anisotropic and thus it can not be described by scalar transport coefficients but by shear - rate dependent tensorial quantities whose explicit determination has been the main objective of this paper . in order to gain some insight into this complex problem ,
a dilute binary mixture of maxwell molecules under usf has been considered .
this is perhaps the only interaction potential for which the boltzmann equation can be exactly solved in some specific non - homogenous situations , such as in the case of the usf problem .
in particular , the corresponding rheological properties of the mixture ( nonlinear shear viscosity and viscometric functions ) have been obtained for _ arbitrary _ values of the shear rate and without any restriction on the parameters of the mixture ( masses , concentrations , and force constants ) .
@xcite this exact solution is of great significance in providing insight into the type of phenomena that can occur in conditions far away from equilibrium . in this paper , the interest has been focused on situations that slightly deviate from the usf by small spatial gradients . under these conditions , a generalized chapman - enskog method @xcite around the shear flow distribution has been used to determine mass transport in the first order of the deviations of the hydrodynamic field gradients from their values in the reference shear flow state @xmath231 . in this case , the mass flux @xmath232 is given by eq .
( [ 3.25 ] ) , where the corresponding set of generalized transport coefficients @xmath233 are the solutions of the coupled algebraic equations ( [ 4.23])([4.25 ] ) in the unthermostatted case , while they are explicitly given by eqs.([4.30])([4.32 ] ) in the presence of a gaussian thermostat .
@xcite this type of external forces are usually employed in nonequilibrium molecular dynamics simulations to compensate exactly for the viscous increase of temperature .
as expected , the results show that the coefficients @xmath233 present a complex dependence on the shear rate and on the masses , mole fractions , and force constants .
this is clearly illustrated in figs .
[ fig1][fig6 ] for the tracer limit case ( @xmath181 ) .
the deviations of @xmath234 from their equilibrium values are basically due to three different reasons .
first , the presence of shear flow modifies the collision frequency of the conventional diffusion problem ( @xmath235 ) by a shear - rate dependent term .
second , given that the binary mixture is in general constituted by particles mechanically different , the reference shear flow states @xmath106 and @xmath116 are completely different .
this effect gives rise to terms proportional to @xmath236 .
third , in the unthermostatted case , the generalized coefficients are coupled due to the inherent non - newtonian features of the usf state .
each one of the three effects is a different reflection of the extreme nonequilibrium conditions present in the mixture .
it is apparent that the results presented here in the particular case of maxwell molecules may be relevant for interpreting computer simulation results .
sarman , evans , and baranyai @xcite carried out time ago molecular dynamics simulations in a strongly sheared lennard - jones binary mixture to evaluate the self- and mutual - diffusion tensor by means of green - kubo formulae .
@xcite to the best of my knowledge , this is the only computer experiment in which the shear - rate dependence of the diffusion tensor @xmath16 has been measured .
they considered an equimolar lennard - jones mixture at two different densities and the parameters in the potential were adjusted to model an argon - krypton mixture , which means that the two components are fairly similar .
as already said in ref . , when one considers this type of mixture ( @xmath237 , @xmath238 , @xmath176 ) in the thermostatted case , the general qualitative dependence of the ( reduced ) mutual diffusion tensor @xmath239 on the ( reduced ) shear rate agrees quite well with computer simulations .
thus , theory and simulation predict that in general , the @xmath240 element increases to a maximum and then it decreases again , while the @xmath209 element decreases with increasing shear rate . the off - diagonal elements @xmath219 and @xmath241 are negative and their magnitude increases with @xmath192 for not very large values of the shear rate .
however , kinetic theory predicts that @xmath242 , while the opposite happens in computer simulations . on the other hand , at a quantitative level , the influence of shear flow on diffusion is much more modest in the molecular dynamics simulations than the one found theoretically for dilute gases .
this is probably due to the fact that the shear rates ( in reduced units ) applied in the simulations are not large enough to observe significant changes of the diffusion tensor relative to its equilibrium value .
an alternative to overcome the difficulties for reaching large shear rates in nonequilibrium molecular dynamics at low - density is the direct simulation monte carlo method .
i hope that the results derived here for maxwell molecules for @xmath16 , @xmath145 and @xmath146 stimulate the performance of monte carlo simulations to assess the reliability of the maxwell results to describe mass transport in strongly sheared mixtures for more realistic interaction potentials . as said before
, it must noted that , in order to observe large effects of shear flow on the tensors @xmath243 , the ( reduced ) shear rate must be at least of the order of 1 .
this means that for the inert gas fluids considered in this paper , non - newtonian effects on mass transport could be observable for shear rates practically unattainable in the laboratory . in this sense
, one should look at fluids that are observed to be non - newtonian , such as colloidal suspensions , polymeric liquids , gels,@xmath244 .
@xcite although the results derived in this paper have been focused on the mass transport , the remaining transport coefficients associated with the pressure tensor @xmath245 and the heat flux vector @xmath246 could be determined from the integral equations ( [ 3.18])([3.21 ] ) .
nevertheless , in practice this calculation can not be carried out analytically by using the boltzmann equation since the fourth - degree moments of usf ( whose explicit expressions are not known in the boltzmann equation , except for a single gas@xcite ) are needed to get the heat flux . in order to overcome such a difficulty one can use a model kinetic equation that preserves the essential features of the true boltzmann equation but admits a more practical analysis .
perhaps the most well - known model for gas mixtures is the gross - krook ( gk ) kinetic model .
@xcite in this model the boltzmann operator @xmath247 $ ] is replaced by the relaxation term @xmath248\to -\nu_{rs}(f_r - f_{rs}),\ ] ] where @xmath249 \label{5.2}\ ] ] and @xmath250 @xmath251 \;.
\label{5.4}\ ] ] the partial temperatures @xmath252 are defined by eq .
( [ 2.5 ] ) . for maxwell molecules ,
the effective collision frequency @xmath253 is given by @xmath254 where @xmath255 is a constant to be fixed by requiring that the model reproduces some transport coefficient of the boltzmann equation .
an exact solution to the gk kinetic model for a binary mixture in usf has been found .
@xcite the comparison of the gk results with those from the boltzmann equation @xcite at the level of the rheological properties shows good agreement , confirming the reliability of the gk model in computing transport properties in far from equilibrium situations as well .
starting from the usf solution of the gk model , the fluxes @xmath245 and @xmath246 are obtained in appendix [ appd ] in the thermostatted case . with all the transport coefficients
known , the constitutive equations for the mass , momentum and heat fluxes are completed and the corresponding set of closed hydrodynamic equations for the mixture can be derived .
this allows one to perform a linear stability analysis of the hydrodynamic equations with respect to the usf and determine the conditions for instabilities at long wavelengths .
previous results for a single gas @xcite have shown that usf is unstable when the perturbations are along the velocity gradient ( @xmath227 direction ) .
the problem now is to extend this analysis to the case of multicomponent systems .
work along this line will be reported in the near future .
partial support from the ministerio de ciencia y tecnologa ( spain ) through grant no .
fis200760977 and from the junta de extremadura through grant no .
gru07046 is acknowledged .
in this appendix , some technical details on the determination of the first - order approximation @xmath120 by means of the chapman enskog - like expansion are provided . inserting the expansions ( [ 3.8 ] ) and ( [ 3.11 ] ) into eq .
( [ 3.1 ] ) , one gets the kinetic equation for @xmath120 : @xmath256f_1^{(1)}.\ ] ] the velocity dependence on the right - hand side of eq .
( [ a1 ] ) can be obtained from the macroscopic balance equations ( [ 3.3])([3.5 ] ) to first order in the gradients .
they are given by @xmath257 @xmath258 @xmath259 @xmath260 where use has been made of the result @xmath261 .
in addition , @xmath262 use of eqs .
( [ a2])([a5 ] ) in eq .
( [ a1 ] ) yields @xmath263 where @xmath264 @xmath265 @xmath266 @xmath267 upon writing eq .
( [ a11 ] ) use has been made of the expression of the total pressure tensor @xmath245 of the mixture @xcite @xmath268 where @xmath269 is the viscosity tensor .
the solution to eq .
( [ a7 ] ) has the form given by eq .
( [ 3.17 ] ) , where the coefficients @xmath142 , @xmath143 , @xmath144 , and @xmath123 are functions of the peculiar velocity and the hydrodynamic fields @xmath1 , @xmath2 , @xmath3 , and @xmath78 . the time derivative acting on these quantities can be evaluated with the replacement @xmath270 moreover , there are contributions from @xmath271 acting on the pressure , temperature , and velocity gradients given by @xmath272 @xmath273 @xmath274 the corresponding integral equations ( [ 3.18])([3.20 ] ) can be obtained when one identifies coefficients of independent gradients in eq .
( [ a7 ] ) and takes into account eqs .
( [ a15])([a17 ] ) and the mathematical property @xmath275 where in the last step it has been taken into account that @xmath276 depends on @xmath78 through @xmath277 .
in the unthermostatted case ( @xmath88 ) , the integral equations defining the generalized transport coefficients @xmath16 , @xmath145 and @xmath146 are given by eqs .
( [ 4.1])([4.3 ] ) . to get these coefficients , one multiplies ( [ 4.1])([4.3 ] ) by @xmath156 and integrates over velocity .
the result is @xmath278 @xmath279\nonumber\\ & & = m_1\int{\text{d}}{\bf c}\ ; c_i b_{1,j}+ \frac{2a\rho}{3p^2}d_{t , ij}(1-p\partial_p)p_{xy}^{(0)},\end{aligned}\ ] ] @xmath280\nonumber\\ & & = m_1\int{\text{d}}{\bf c}\ ; c_i c_{1,j}- \frac{2a\rho}{3p}d_{p , ij}(\partial_tp_{xy}^{(0)}),\end{aligned}\ ] ] where @xmath281 upon writing eqs .
( [ 4.7])([4.9 ] ) , use has been made of the relation ( [ 2.10.1 ] ) , which yields the results @xmath282 @xmath283 @xmath284 the velocity integrals appearing in eqs .
( [ 4.7])([4.9 ] ) can be performed by using eqs .
( [ a8])([a10 ] ) , @xmath285 @xmath286 @xmath287 the generalized transport coefficients @xmath16 , @xmath145 , and @xmath146 can be written as @xmath288 , @xmath289 , and @xmath290 where @xmath211 , @xmath212 , and @xmath213 are dimensionless functions of the shear rate . moreover , from dimensional analysis , @xmath291 , @xmath292 , and @xmath293 .
therefore , @xmath294
@xmath295 @xmath296 where use has been made of the identity @xmath297 with @xmath298 .
taking into account the above results one arrives at the set of algebraic equations ( [ 4.23])([4.25 ] ) .
the explicit expressions for the pressure tensors @xmath299 in the usf are provided in this appendix for the special case of tracer limit ( @xmath181 ) .
the nonzero elements of @xmath300 are given by @xcite @xmath301 @xmath302 @xmath303 where @xmath67 , @xmath304 , @xmath68 being defined by eq .
( [ 4.4 ] ) .
moreover , @xmath305 where @xmath69 is the mass ratio . the ( reduced ) thermostat parameter is given by @xmath306 where @xcite @xmath307 where @xmath308 $ ] and @xmath309 .
usually , @xmath310 except for very large shear rates and/or very disparate mass binary mixtures .
@xcite the nonzero elements of @xmath311 are given by @xcite @xmath312 @xmath313\right\},\end{aligned}\ ] ] @xmath314 where @xmath315 , @xmath316 , and @xmath317
this appendix addresses the evaluation of the fluxes @xmath245 and @xmath246 in the thermosttated case from the gk kinetic model ( [ 5.1 ] ) .
the first order corrections to the fluxes are @xmath318 @xmath319 where the partial contributions to the transport coefficients are defined as @xmath320 @xmath321 @xmath322 @xmath323 from the above partial contributions one can get the generalized shear viscosity @xmath324 , the generalized duffour coefficient @xmath325 , the generalized pressure energy coefficient @xmath326 , and the generalized thermal conductivity @xmath327 .
the quantities @xmath328 still verify the integral equations ( [ 3.18])([3.21 ] ) ( with @xmath111 ) with the only replacement @xmath329 where @xmath330 , @xmath331 and @xmath332 in eq .
( [ d8 ] ) , @xmath333 . in order to get the coefficients @xmath334 , @xmath335 , @xmath336 and @xmath337 , it is convenient to introduce the velocity moments @xmath338 @xmath339 @xmath340 @xmath341 and similar definitions for the species @xmath342 .
the knowledge of the above moments allows one to get the expressions of the coefficients @xmath343 , @xmath344 , @xmath345 and @xmath346 .
the method to evaluate the moments @xmath347 , @xmath348 , @xmath349 , and @xmath350 is quite similar . here , as an example ,
the mathematical steps to determine the moments @xmath347 associated with the transport coefficients @xmath344 will be analyzed in detail .
first , in the thermosttated case , eq . ( [ 3.18 ] ) with the change ( [ d7 ] ) becomes @xmath351 where @xmath352 is given by eq .
( [ a8 ] ) . upon writing ( [ d14 ] ) use has been made of the constitutive form ( [ 3.25 ] ) for the mass flux .
now , we multiply eq .
( [ d14 ] ) by @xmath353 and integrate over velocity .
after some algebra , we get @xmath354 x_{k,\ell , m}^{(i)}=r_{k,\ell , m}^{(i)},\ ] ] where @xmath355\nonumber\\ & & \times \lambda_{k+\delta_{jx},\ell+\delta_{jy},m+\delta_{jz}}d_{ji},\end{aligned}\ ] ] @xmath356 in eqs .
( [ d16 ] ) and ( [ d17 ] ) , we have introduced the temperature ratios @xmath357 and @xmath358 and the ( unperturbed ) moments of the usf @xmath359 the explicit shear - rate dependence of @xmath360 , @xmath361 and @xmath362 can be found in ref .. moreover , @xmath363 if @xmath364 are even , being zero otherwise .
the solution to eq .
( [ d15 ] ) can be written as @xmath365^{-(1+q)}r_{k - q,\ell + q , m}^{(i)}\;. \label{d20}\ ] ] note that eq .
( [ d20 ] ) is still formal since one needs to know the coefficients @xmath16 .
they can be consistently determined from their definitions ( [ 3.26 ] ) . once these coefficients are known , eq .
( [ d20 ] ) allows one to get the coefficients @xmath344 .
the same method can be applied to evaluate the remaining moments .
the moments @xmath366 and @xmath367 are given by @xmath368^{-(1+q)}s_{k - q,\ell + q , m}^{(i)}\ ; , \label{d21}\ ] ] @xmath369^{-(1+q ) } t_{k - q,\ell+q , m}^{(i)}\ ; , \label{d21.1}\ ] ] where @xmath370\nonumber\\ & & \times \lambda_{k+\delta_{jx},\ell+\delta_{jy},m+\delta_{jz}}d_{p , ji},\end{aligned}\ ] ] @xmath371\nonumber\\ & & \times \lambda_{k+\delta_{jx},\ell+\delta_{jy},m+\delta_{jz}}d_{t , ji}.\end{aligned}\ ] ] the expressions of @xmath372 and @xmath373 are formally identical to that of @xmath374 , except that the operator @xmath375 appearing in ( [ d17 ] ) must be replaced by the operators @xmath376 and @xmath377 in the cases of @xmath372 and @xmath373 , respectively .
finally , the expression of @xmath378 is @xmath379^{-(1+q ) } \left [ u_{k - q,\ell+q , m}^{(ij)}+a\delta_{iy}w_{k,\ell , m}^{(xj)}\right],\ ] ] where @xmath380 the transport coefficients @xmath344 , @xmath345 , @xmath346 , and @xmath343 can be obtained from eqs .
( [ d20 ] ) , ( [ d21 ] ) , ( [ d21.1 ] ) and ( [ d22bis ] ) , respectively , in terms of the shear rate and the parameters of the mixture . their respective counterparts for species @xmath342 can be easily determined from them by making the changes : @xmath381 , @xmath382 , and @xmath383 . the expression of the duffour tensor @xmath384 coincides with the one obtained before @xcite in a stationary state with @xmath385 .
finally , note that the usf moments @xmath362 can be written as @xcite @xmath386 where the dimensionless moments @xmath387 depend on @xmath2 and @xmath3 through their dependence on @xmath192 .
consequently , @xmath388 @xmath389 | transport coefficients associated with the mass flux of a binary mixture of maxwell molecules under uniform shear flow are exactly determined from the boltzmann kinetic equation .
a normal solution is obtained via a chapman enskog - like expansion around a local shear flow distribution that retains all the hydrodynamics orders in the shear rate . in the first order of the expansion
the mass flux is proportional to the gradients of mole fraction , pressure , and temperature but , due to the anisotropy induced in the system by the shear flow , mutual diffusion , pressure diffusion and thermal diffusion tensors are identified instead of the conventional scalar coefficients .
these tensors are obtained in terms of the shear rate and the parameters of the mixture ( particle masses , concentrations , and force constants ) .
the description is made both in the absence and in the presence of an external thermostat introduced in computer simulations to compensate for the viscous heating .
as expected , the analysis shows that there is not a simple relationship between the results with and without the thermostat .
the dependence of the three diffusion tensors on the shear rate is illustrated in the tracer limit case , the results showing that the deviation of the generalized transport coefficients from their equilibrium forms is in general quite important .
finally , the generalized transport coefficients associated with the momentum and heat transport are evaluated from a model kinetic equation of the boltzmann equation . |
violation of @xmath20 symmetry was observed in the difference of @xmath22 decay and @xmath23 decay , where data with @xmath24 chosen to be @xmath25 or @xmath26 @xcite . in the experiment of producing @xmath27
entangled state in the @xmath28 resonance decay , conditions of @xmath29 or @xmath30 decay into @xmath31 or @xmath32 and another @xmath29 or @xmath30 decay into @xmath33 or @xmath34 were chosen .
we consider violation of @xmath20 symmetry or time reversal symmetry in the decay of @xmath35 mesons observed in the difference of @xmath36 where @xmath37 and @xmath38 are linear combinations of @xmath15 and @xmath18 , @xmath39 where @xmath40 . from the difference of experimental data of @xmath41
one can evaluate the ckm angle @xmath42 and @xmath20 violation@xcite , but we do not take this approach , since an entangled neutral @xmath35 meson state @xmath28 affects the phase @xmath42@xcite .
cartan s supersymmetry@xcite defines the interaction of spinors ( quarks or anti - quarks in @xmath29 meson ) and vector particles ( @xmath43 bosons or @xmath42 particles ) . in our application of the model to the electro - magnetic decay of a higgs particle ( @xmath44 ) and weak decay ( @xmath45 )
@xcite suggest that the quark - gluon and the quark - photon interaction of cartan s supersymmetry could give clear signal of violation of the @xmath20 symmetry . in the standard model ,
@xmath20 symmetry violation occurs from the interference of the tree diagram amplitude and the penguin diagram amplitude@xcite .
typical tree diagrams and penguin diagrams of this decay are shown in fig.[b0kt ] and fig.[b0kp ] . decay . in our model
we include the loop of @xmath42s in addition to that of @xmath5 bosons.,width=226 ] decay . in our model
we include the loop of @xmath42s in addition to that of @xmath5 bosons.,width=226 ] the decay rates of @xmath46 is parametrized as @xmath47\},\nonumber\ ] ] where @xmath48 is the difference of the decay time to the @xmath20 eigenstates and the decay time to the flavor eigenstates , and @xmath49 is the mass difference between the physical states of the neutral @xmath35 meson systems .
the decay time difference @xmath50 is approximately given by the distance @xmath51 between the decay points and the lorentz boost @xmath52 as @xmath53 . adopting the parametrization of the low mass eigenstate @xmath54 and the high mass eigenstate @xmath55 as @xmath56 and @xmath57 , if symmetry holds .
the amplitude @xmath58 and @xmath59 are parametrized as @xmath60 where using the @xmath20 parity @xmath61 for @xmath62 and @xmath63 here @xmath64 is the ratio of the @xmath65 and @xmath66 decays .
experimentally , violation of @xmath20 or time reversal symmetry can be observed via difference of @xmath50 dependence of @xmath29 tagged events and @xmath30 tagged events@xcite .
the difference of @xmath29 tagged events and @xmath30 tagged events shows the direction of the time axis@xcite .
the propagator between vertices @xmath67 and @xmath68 is assigned as a @xmath5 boson , and the source of @xmath69 in the penguin diagram is assigned as a @xmath42 photon or @xmath70 boson . in our model , a @xmath42 particle @xmath71 is emitted from a quark in @xmath29 or @xmath30 as @xmath72 the quark @xmath73 which consists of @xmath74 and its small component @xmath75 interacts with an anti - quark @xmath76 which consists of @xmath77 and its small component @xmath78 .
a @xmath42 particle @xmath71 is emitted also from @xmath76 as @xmath79 in the @xmath80 theory , @xmath71 , @xmath81 and @xmath82 are related to the gauge fields @xmath83 and @xmath84 by @xmath85 where @xmath86 is the weinberg angle . in the standard model , the weak interaction
is described as@xcite @xmath87 where @xmath88 and @xmath89 are right - handed spinors .
the weak isospin of quarks or leptons are defined by @xmath90 .
the isospin of @xmath91 and @xmath92 are rotated , and physical states @xmath82 , @xmath71 , @xmath81 and @xmath73 appear .
we expect that physical @xmath76 appear together with the physical @xmath73 .
the cabibbo - kobayashi - maskawa ( ckm ) amplitudes @xmath93 are parametrized as @xmath94\nonumber\end{aligned}\ ] ] where @xmath95 .
large @xmath96 yields strong @xmath97 decay modes .
the decay mode @xmath98 is strong due to relatively large @xmath99 and small @xmath100 and @xmath101 .
the relation @xmath102 requires @xmath103 , while @xmath96 requires @xmath104 , and the parametrization of the ckm amplitudes are not completed .
we study also the decay of @xmath105 and @xmath106 . in section 2 , we explain cartan s supersymmetry in @xmath107 interaction , and in section3 , we explain differences of @xmath108 and @xmath109 from cartan s supersymmetry .
we want to reproduce the qualitative features of ckm amplitudes from the trialty model . in the previous work@xcite , we could understand absence of @xmath110 , and presence of @xmath111 from the triality selection rules that cartan s supersymmetry predicts@xcite .
we define a dirac spinor composed of @xmath74 and @xmath75 @xmath112 and another dirac spinor composed of @xmath77 and @xmath78 @xmath113 which interact with four dimensional vector fields @xmath24 .
the trilinear form of electromagnetic interaction in these bases is @xmath114 in the case of weak interaction , we replace the coupling @xmath115 to @xmath116 and try to make the couplings between fermions and vector particles become unified in the form @xmath117 by suitable choice of @xmath118 or @xmath119 . except the term @xmath120 , which is @xmath121 type , it is possible by the following choice @xmath122 we allow also direct channel couplings of @xmath123 and @xmath124 .
experimentally , violation of @xmath20 was observed in the decay of @xmath29 and @xmath125 around 1997 . in the case of @xmath126 , there is a constraint on the unitarity angle @xmath42 , which is called fleischer - mannel bound by grossmann et al@xcite .
they derived the bound for the angle @xmath42 of unitary triangle , @xmath127 , with @xmath128 and the uncertainty @xmath129 . in comparison to @xmath126 decay
, @xmath130 contains weak final state interactions . in order to evaluate the @xmath130 decay
amplitudes we fix the initial configuration of @xmath131 and the final configuration @xmath132 , @xmath133 , and a quark is expressed as a dirac spinor @xmath134 and an anti - quark is expressed as a dirac spinor @xmath135 . in penguin diagrams of @xmath130 ,
an anti - quark @xmath136 emits a vector particle @xmath137 and transforms itself to a @xmath138 quark , emits a vector particle @xmath137 that changes to a @xmath25 , and absorbs the vector particle and transforms itself to a @xmath16 quark .
couplings of loops of a vector particle and an antiquark can be @xmath139 type , shown in fig.[bscc11 ] and @xmath140 type , shown in fig.[bscc55 ]
. we can combine the @xmath140 and the @xmath139 type to make the effective coupling @xmath141 type .
decay , penguin diagram @xmath140 type . in the @xmath25 configurations
, @xmath142 decays to @xmath143 , @xmath144 or @xmath145.,width=226 ] [ bscc55b ] decay , penguin diagram @xmath140 type . in the @xmath25 configurations
, @xmath142 decays to @xmath143 , @xmath144 or @xmath145.,width=226 ] [ bscc55a1 ] decay , penguin diagram @xmath140 type . in the @xmath25 configurations
, @xmath142 decays to @xmath143 , @xmath144 or @xmath145.,width=226 ] [ bscc55a ] decay , penguin diagram @xmath140 type . in the @xmath25 configurations
, @xmath142 decays to @xmath143 , @xmath144 or @xmath145.,width=226 ] decay , penguin diagram @xmath140 type . in the @xmath25 configurations
, @xmath142 decays to @xmath143 , @xmath144 or @xmath145.,width=226 ] [ bscc551234_4 ] decay , penguin diagram @xmath140 type . in the @xmath25 configurations
, @xmath142 decays to @xmath143 , @xmath144 or @xmath145.,width=226 ] [ bscc5512_3 ] decay , penguin diagram @xmath140 type . in the @xmath25 configurations
, @xmath142 decays to @xmath143 , @xmath144 or @xmath145.,width=226 ] [ bscc5531_2 ] decay , penguin diagram @xmath140 type . in the @xmath25 configurations
, @xmath142 decays to @xmath143 , @xmath144 or @xmath145.,width=226 ] in penguin diagrams of @xmath146 , a quark @xmath147 emits a vector particle @xmath137 and transforms itself to a @xmath148 quark , and emits a vector particle @xmath137 that changes to a @xmath25 , and absorbs the vector particle and transforms itself to an @xmath19 quark .
couplings of a loop of vector particles and an antiquark can be @xmath139 type , shown in fig.[bscc11_14 ] and @xmath140 type , shown in fig.[bscc55_14 ] .
the coupling of a @xmath148 quark and a vector particle @xmath137 is pseudoscalar , which we indicate by @xmath149 .
[ bscc123_0 ] [ bscc33_34 ] decay , penguin diagram @xmath140 type . in the @xmath25 final state
, @xmath142 decays to @xmath143 , @xmath144 or @xmath145.,width=226 ] [ bscc22_24 ] decay , penguin diagram @xmath140 type . in the @xmath25 final state
, @xmath142 decays to @xmath143 , @xmath144 or @xmath145.,width=226 ] decay , penguin diagram @xmath140 type . in the @xmath25 final state
, @xmath142 decays to @xmath143 , @xmath144 or @xmath145.,width=226 ] [ bscc11_1234 ] decay , penguin diagram @xmath140 type . in the @xmath25 final state
, @xmath142 decays to @xmath143 , @xmath144 or @xmath145.,width=226 ] [ bscc11_124 ] decay , penguin diagram @xmath140 type . in the @xmath25 final state
, @xmath142 decays to @xmath143 , @xmath144 or @xmath145.,width=226 ] [ bscc11_314 ] decay , penguin diagram @xmath140 type . in the @xmath25 final state
, @xmath142 decays to @xmath143 , @xmath144 or @xmath145.,width=226 ] the penguin diagrams of @xmath17 of @xmath150 type are simple charge conjugation of @xmath107 , but those of @xmath13 type are different .
space components of the @xmath136 quark in @xmath29 are small components of dirac spinors , while those of the @xmath147 quark in @xmath30 are large components of spinors .
it means that amplitudes of @xmath13 contribution in @xmath14 are suppressed as compared to @xmath17 . in the @xmath151 decay , there are tree diagrams of type i , in which vector particle of @xmath69 is emitted from @xmath136 , or tree diagrams of type ii , in which a vector particle that decays to @xmath152 is emitter from @xmath16 , and after final state interactions @xmath25 and @xmath15 are created . in the analysis of coulomb interaction between electrons and holes using quaternion basis@xcite , it was necessary to make couplings of a photon @xmath153 to an electron and a photon @xmath142 to electrons equivalent , for reproducing electron - hole interactions . in the present case , couplings of a vector particle @xmath137 produced from a @xmath154 pair of @xmath25 to a @xmath155 of @xmath156 , and couplings of a vector particle @xmath153 produced from a @xmath157 pair in @xmath25 and a @xmath155 in @xmath15 are equivalent , and we can create a @xmath158 or @xmath159 vertex of @xmath153 from the source of @xmath137 . by the final state interaction , the created @xmath148 and @xmath138 interchanged from @xmath136 in @xmath29
produce @xmath25 meson , and the created @xmath19 quark and the original @xmath160 quark of @xmath29 produce a @xmath18 meson . in the evaluation of the matrix elements , we choose the original quark and the original anti - quarks are large components , if it is possible .
the time component of the amplitude of @xmath14 and @xmath161 of @xmath140 type contains the small component @xmath162 and they are expected to be much smaller than those of the @xmath139 type .
decay ( type ii ) .
the @xmath25 is not produced from the vector particle @xmath137 .
, width=226 ] decay ( type ii ) .
the @xmath25 is not produced from the vector particle @xmath137 .
, width=226 ] decay ( type ii ) .
the @xmath25 is not produced from the vector particle @xmath137 .
, width=226 ] decay ( type ii ) .
the @xmath25 is not produced from the vector particle @xmath137 .
, width=226 ] decay ( type ii ) .
the @xmath25 is not produced from the vector particle @xmath137 .
, width=226 ] decay ( type ii ) .
the @xmath25 is not produced from the vector particle @xmath137 .
, width=226 ] decay ( type ii ) .
the @xmath25 is not produced from the vector particle @xmath137 .
, width=226 ] decay ( type ii ) .
the @xmath25 is not produced from the vector particle @xmath137 .
, width=226 ] in the case of @xmath139 type interactions , diagrams of type i is symmetric under charge conjugation , but diagrams of type ii @xmath163 and @xmath164 as shown in fig.[bdbart15_44 ] are not symmetric .
the @xmath139 interaction of @xmath163 , @xmath25 consists of @xmath165 and @xmath15 consists of @xmath166 , i.e. @xmath15 has the small component @xmath162 , or the strength of the tree diagram of @xmath167 becomes weak , same as in penguin diagram of @xmath140 type and the number of difference of events @xmath168 becomes large . in the case of space components of @xmath163 , suppression of @xmath25 creation occurs , but it would be difficult to detect from experimental data of @xmath167 .
the @xmath139 interaction of @xmath164 , @xmath25 consists of @xmath169 , and @xmath18 consists of @xmath170 , i.e. @xmath25 has the small component @xmath162 .
effects of this suppression on the strength of the @xmath171 may be difficult to detect , but in the case of the space components of @xmath164 , suppression of @xmath18 creation occurs , and becomes weak , same as in penguin diagrams of @xmath13 type , and the number of difference of events @xmath168 becomes small .
the @xmath29 tags and the @xmath30 tags in the above properties are what the babar experiment of @xmath172 of @xmath29 tags and @xmath30 tags in the @xmath173 region show @xcite .
decay ( left ) and @xmath174 decay ( right ) .
, width=226 ] decay ( left ) and @xmath174 decay ( right ) .
, width=226 ] the experimental data of @xmath20 even final states @xcite shows enhancements in raw @xmath20 asymmetry in @xmath175ps region .
we expect this is the contribution of type ii tree diagram generating @xmath176 , and that of @xmath177 type penguin diagrams generating @xmath178 which is stronger than @xmath179 . in our @xmath35 decay to @xmath180 ,
the cp - odd final state of @xcite can be approximated by @xmath181))^2-(e^{-(t+1)/16}(1+\cos[\delta m_{b } ( t+2)]))^2\nonumber\\ & & = e^{(t-1)/8 } 4\cos^4\frac{\delta m_b(t-2)}{2}-e^{-(t+1)/8 } 4\cos^4\frac{\delta m_b(t+2)}{2}\end{aligned}\ ] ] and cp - even final state of @xcite can be approximated by @xmath182))^2-\frac{1}{8 } ( e^{-(t+1)/16}(1-\cos[{\delta m_{b}}(t+2)]))^2\nonumber\\ & & = e^{(t-1)/8 } \frac{1}{2}\sin^4\frac{\delta m_b(t-2)}{2}-e^{-(t+1)/8 } \frac{1}{2}\sin^4\frac{\delta m_b(t+2)}{2}\end{aligned}\ ] ] where @xmath183ps@xmath184 is fixed@xcite , which are shown in fig.[cpasymmetry ] .
center of an elliptic disk is that of an experimental point , and the error bar of the asymmetrys are fixed to 0.2 , and the ratio of error bars in asymmetry axis and in @xmath50 axis is fixed to the goldenratio @xmath185 which ratio is the value used in mathematica for presentations of asymmetry curves .
decay of cp - odd final state ( left ) and cp - even final state ( right ) .
, width=226 ] decay of cp - odd final state ( left ) and cp - even final state ( right ) .
, width=226 ] we expect that the experimental asymmetry of @xmath186 and @xmath17 at large @xmath50 can be understood from the final state wave functions caused by the @xmath187 vertexes .
cartan s supersymmetry based on clifford algebra presents large asymmetry between the weak decay of fermions @xmath73 and @xmath191 .
the origin of the discrepancy between the raw asymmetry of @xmath20 even final states @xmath192 and the best fit projection in @xmath50 of @xmath29 tagged events and @xmath30 tagged events@xcite is expected to be due to the @xmath140 type penguin interaction .
one can ask whether similar effects can be seen in other decay processes leading to different final states .
there are analysis of @xmath20 symmetry in @xmath193 , @xmath194 and @xmath195@xcite . in the analysis of @xmath193 , penguin diagrams of @xmath13
were not included and the @xmath20 asymmetry was not observed .
results of @xmath193@xcite is less clear than that of @xmath196 . in the penguin diagram of @xmath197 ,
the vector particle @xmath69 emitted from the anti - quark transformed to @xmath25 , while in the case of @xmath108 , the @xmath148 is absorbed in @xmath198 and @xmath138 is absorbed in @xmath199 .
there is not large differences in penguin diagrams and tree diagrams , and clear @xmath20 asymmetry was not observed .
an experiment of @xmath200 using partial reconstruction of @xmath201 and a kaon tag@xcite does not show clean @xmath20 asymmetry .
decay.,width=226 ] decay.,width=226 ] decay.,width=226 ] decay.,width=226 ] decay.,width=226 ] decay.,width=226 ] decay.,width=226 ] decay.,width=226 ]
we applied cartan s supersymmetry to weak interaction and analyzed the @xmath107 and @xmath202 and @xmath203 decay using vector particles @xmath1 and @xmath204 ( @xmath3 ) and assigning couplings to quarks by @xmath187 vertices , and inducing effective couplings of @xmath5 bosons , @xmath70 bosons and photons .
we observed that the @xmath20 symmetry of @xmath205 is violated , while that of @xmath206 and @xmath195 are not strongly violated . in these decay processes ,
there appear penguin diagrams and tree diagrams , and in the @xmath20 asymmetry of @xmath205 and @xmath207 , space components of penguin diagrams of @xmath140 type and the time component of tree diagrams of @xmath13 type play important roles in changing the strength of @xmath167 events and @xmath208 events , where @xmath69 is not necessarily the @xmath25 state .
we showed that cartan s supersymmetry is a useful tool for analyzing violation of @xmath20 symmetry of @xmath209 and @xmath210 systems .
difficulties in choosing appropriate ckm amplitudes @xmath211 and @xmath212 that explain @xmath213 and @xmath214 consistently can be removed by an appropriate assignment of the strength of penguin diagram amplitudes and tree diagram amplitudes @xcite .
29 j. bernabeu and f. martinez - vidal , _ colloquium _ : time - reversal violation with quantum - entangled @xmath35 mesons , * 87 * ( 2015 ) 165 - 182 .
b. aubert et al , ( babar collaboration ) , * 79 * ( 2009 ) 072009 ; arxiv:0902.1708v2[hep - ex ] r. fleischer and t. mannel , constraining the ckm angle @xmath42 and penguin contributions through combined @xmath215 branching ratios , [ hep - ph/9704423 ] .
y. grossman , y. nir , s. plaszezynski and m - h .
schune , implication of the fleisher - mannel bound , [ hep - ph/9709288 v1 ] .
cartan , _ la thorie des spineurs _ ( 1937 ) , _ the theory of spinors _ , dover pub .
( 1966 ) p.118 .
s. furui , fermion flavors in quaternion basis and infrared qcd , few body syst .
* 52 * , ( 2012 ) 171 - 187 , [ hep - ph/1104.1225 ] .
s. furui , triality selection rules of octonion and quantum mechanics , [ hep - ph/1409.3761 ] .
s.furui , cartan s supersymmetry and weak and electromagnetic interactions , [ hep - ph/1502.04524 ] s. furui , cartan s supersymmetry and the decay of a @xmath216 , [ hep - ph/1504.03795 ] .
m. zeller , particle decays point an arrow of time , physica * 5 * ( 2012 ) 129 - 130 .
p. becher , m , bhm and h. joos , _ eichtheorien _ , teubener studienbcher ( 1981 ) p.347 .
s. furui and t. takano , on the amplitude of external perturbation and chaos via devil s staircase - stability of attractors - , submitted to nonlinear dynamics ; arxiv:1312.3991[nlin.cd ] s. furui , cartan s supersymmetry and the decay of @xmath217 to @xmath218 and @xmath219 , abstract of fpcp2015 , nagoya university ( 2015 ) .
r. fleischer , extracting @xmath42 from @xmath220 and @xmath221 , hep - ph/9903455 .
k. de bruyn and r. fleischer , a roadmap to cintrol penguin effects in @xmath222 and @xmath223 , 1412.06834 v3 [ hep - ph ] .
r. fleischer , towards new frontiers in @xmath20 violation in b decays , journal of physics : conference series * 556 * ( 2014 ) 012035 , iop publishing .
b. aubert et al , ( babar collaboration ) , direct @xmath20 asymmetry in @xmath224 decays , * 93 * ( 2004 ) 131801 .
lees et al , ( babar collaboration ) , search for @xmath20 violation in @xmath225 mixing using partial reconstruction of @xmath226 and a kaon tag , * 111 * ( 2013 ) 101802 .
s. furui , departure from the standard model of meson decays and cartan s supersymmetry , arxiv:1511.07774[physics.ge - ph ] . | cartan s supersymmetriy fixes couplings of two types of fermions @xmath0 and two types of vector fields @xmath1 and @xmath2 , ( @xmath3 ) .
the electromagnetic interaction of leptons and quarks is expressed as @xmath4 . in the case of coupling of leptons and quarks with @xmath5 ,
we extend the coupling @xmath6 to @xmath7 , where @xmath8 or @xmath9 , and unify the interactions in the form @xmath10 , where @xmath11 implies appropriate @xmath1 or @xmath12 is chosen .
the @xmath13 term induces in time component of @xmath14 , @xmath15 with the @xmath16 quark dominated by the small component , while in space components of @xmath17 , @xmath18 with the @xmath19 quark dominated by the small component .
this asymmetry could be the origin of the @xmath20 asymmetry in the @xmath14 .
we apply the model to @xmath21 decays and study violation of @xmath20 symmetry via interference of tree and penguin contributions , and observe that the asymmetry is weak consistent to the experiment . |
iron - arsenic - based high - temperature superconductors @xcite encompass a large number of materials with remarkably similar phase diagrams , despite significant differences in their chemical compositions .
both elemental substitutions and external parameters can be used to tune their physical properties.@xcite for example , in bafe@xmath0as@xmath0 there is a high - temperature tetragonal to low - temperature orthorhombic structural phase transition at @xmath2 k accompanied by an antiferromagnetic spin density wave ( sdw ) transition .
the sdw ground state can be suppressed in a number of ways , which in many cases leads to an emergence of superconductivity.@xcite these fe - based superconductors crystallize in the thcr@xmath0si@xmath0-type structure that consists of a square lattice of iron ions sandwiched by pnictogen or chalcogen layers forming tetrahedral chemical bonds with fe .
iron fe@xmath3 ions play an important role in the supercondcutivity , as the main contribution to the density of states near the fermi energy is from the fe @xmath4 orbitals.@xcite the fermi surface ( fs ) of bafe@xmath0as@xmath0 consists of hole and electron pockets at the center ( @xmath5-point ) and at the corner ( @xmath6-point ) of the brillouin zone ( bz ) , respectively , of the body - centered - tetragonal direct lattice containing two formula unit ( f.u . ) .
there are hole and electron pockets that have very similar diameters , which leads to nesting that is most likely responsible for the sdw ordering@xcite and antiferromagnetism in the _ a_fe@xmath0as@xmath0 ( _ a _ = ca , sr , ba , eu ) parent compounds.@xcite previous studies demonstrated that an antiferromagnetic sdw phase is the ground state of these materials and leads to a substantial reconstruction of the fermi surface at low temperature,@xcite which disappears close to the carrier concentration at which the onset of the superconducting dome begins.@xcite the co - substituted bafe@xmath0as@xmath0 compounds are of particular interest since a rigid - band - like change of the chemical potential is reported with co concentration.@xcite .
it has been shown , using angle - resolved photoemission spectroscopy ( arpes ) , that the co substitutions induces a lifshitz transition at the onset of the superconducting dome@xcite and another change in the fs topology occurs with an electron pocket appearing around the @xmath5 point at higher ( @xmath7 ) substitution levels,@xcite which is due to the change in the chemical potential that occurs by providing extra @xmath4 electrons to the system . the large variation of properties of iron arsenic superconductors upon elemental substitutions calls for systematic investigations of the closely related materials , where all fe atoms are substituted by other transition metals .
knowledge of their properties and electronic structure will not only improve understanding of the iron arsenic superconductors , but may also result in discoveries of novel materials with unusual and useful properties .
in particular , the success in synthesizing co - based non - superconducting materials _
a_co@xmath0as@xmath0 ( _ a _ = ca , sr ,
ba , eu ) have attracted significant attention because of their many interesting physical properties.@xcite baco@xmath0as@xmath0 exhibits paramagnetic behavior where the absence of magnetic ordering is considered to be caused by quantum fluctuations.@xcite indeed , this compound is suggested to be a highly mass - renormalized metal near a ferromagnetic quantum critical point.@xcite it is believed that antiferromagnetic spin fluctuations associated with a sdw are the pairing glue for superconductivity in the iron arsenides .
however , ferromagnetic spin fluctuations are strongly pair breaking for spin - singlet pairing .
the calculated electronic density of states ( dos ) of baco@xmath0as@xmath0 at @xmath8 is similar to those of the feas - based compounds ; however , the extra @xmath4 electron in co@xmath3 relative to fe@xmath3 shifts @xmath8 upwards in energy.@xcite this leads to the presence of a sharp peak in the dos at @xmath8 , which is derived from the co @xmath4 orbitals.@xcite therefore , the electronic structure of baco@xmath0as@xmath0 appears very different at @xmath8 than for the _ a_fe@xmath0as@xmath0 compounds .
it is of great interest to clarify the band structure of baco@xmath0as@xmath0 near @xmath8 , which can provide important insights into the microscopic mechanism of the peculiar magnetic and transport properties as of the _ a_co@xmath0as@xmath0 compounds .
so far there are very few studies of this very interesting material @xcite . in this paper
, we present arpes measurements on single crystals of baco@xmath0as@xmath0@xcite that were carried out to clarify the above issues .
we also present lda band structure calculations that are compared with the arpes data and we find reasonable agreement , confirming previous report@xcite .
the measured intensity contours at higher binding energy ( @xmath9 500 mev ) very closely resemble the features near @xmath8 in bafe@xmath0as@xmath0 , which demonstrates that replacing fe ions with co ions causes a nearly rigid shift in the fermi energy by adding two electrons per ( f.u . ) for limited range of binding energies .
we also report the presence of relatively flat bands in the proximity of @xmath8 , that are most likely responsible for peak in dos ( @xmath8 )
. this may be important for understanding the transport and thermodynamic properties of this material .
single crystals of baco@xmath0as@xmath0 were grown out of coas self - flux.@xcite the chemical compositions were checked by wavelength - dispersive x - ray spectroscopy ( wds ) that revealed the desired stoichiometry of @xmath10 .
a rietveld refinement of x - ray diffraction data for crushed crystals confirmed that baco@xmath0as@xmath0 crystallizes in the body - centered tetragonal thcr@xmath0si@xmath0 structure with lattice parameters @xmath11 and @xmath12 at room temperature , in good agreement with literature values.@xcite we have performed high - resolution angle - resolved photoemission spectroscopy ( arpes ) measurements using a scienta r4000 electron analyzer at beamline 7.0.1 at the advanced light source ( als ) , berkeley , california .
all samples were cleaved _ in situ _ , yielding flat mirror - like surfaces in the _ a - b _ plane .
all the arpes data were collected in ultrahigh vacuum below @xmath13 mbar . the energy and momentum resolution were set to @xmath14 mev and @xmath15 , respectively .
we measured several samples , which yielded similar results for the fermi surfaces .
the @xmath8 was determined from the measurements on a au reference sample .
the full - potential linearized augmented - plane - wave ( fp - lapw ) method with the local density approximation@xcite was used to calculate the theoretical fs and band dispersions . to obtain self - consistent charge density
, we employed @xmath16 with muffin tin ( mt ) radii of 2.5 , 2.1 , and 2.1 a.u . for ba , co , and as , respectively .
828 @xmath17-points were selected in the irreducible brillouin zone and calculations were iterated to reach the total energy convergence criterion which was 0.01 mry / cell .
the as atom _ z_-axis position was relaxed to obtain a minimum total energy , yielding _
for the fermi surface calculation , we divided the @xmath19 , @xmath20 ranges of @xmath21 , @xmath22 planes with different @xmath23 values into 200@xmath24200 meshes .
and the theoretical fp - lapw calculations .
the box ( dashed line ) marks the boundary of the first brillouin zone .
arrows indicate the locations of cuts for which the data are plotted in figs 3 and 4 .
( a ) experimental fs of baco@xmath0as@xmath0 at @xmath25 k measured with 85 ev photon energy .
the intensity of the photoelectrons is integrated over @xmath2610 mev about @xmath8 .
( b ) the calculated fermi surface at @xmath27.,width=384 ] as@xmath0 measured with 135 ev photon energy and at @xmath28 k. the intensity of the photoelectrons is integrated over @xmath2610 mev about @xmath8 .
( b ) the calculated fs at @xmath29.,width=384 ] as@xmath0 , measured with 85 ev photon energy and at @xmath25 k. the locations of the cuts # 1 , # 2 , # 3 and # 4 in @xmath30 space are marked in fig .
( lower panels ) calculated band dispersions at @xmath31 .
the white bar at the top of cut # 3 shows the range of momenta for which edcs are plotted in fig .
4.,width=576 ] figure 1(a ) shows the photoemission intensity map at @xmath8 of baco@xmath0as@xmath0 measured with 85 ev photon energy at temperature 90 k. the map is plotted as a function of _
k_@xmath32 , _ k_@xmath33 wave vector components and the intensity is obtained by integrating the spectra within @xmath2610 mev of @xmath8 .
the shapes of the fs sheets in baco@xmath0as@xmath0 are very different from the roughly circular fss present in feas - based materials , which is not surprising , given that co contributes one electron more than fe .
there is a significant difference in the shape of the intensity pattern between the first and second brillouin zones , most likely due to a combination of matrix element effects and three dimensionality of the electronic structure . since arpes can not measure the magnitude of the @xmath34-component of the momentum , we use a comparison with calculations to determine the relation between @xmath23 and photon energy . in fig .
1(b ) we plot the calculated band structure that best matches the data measured with 85 ev photons .
the corresponding value of @xmath23 in the calculations is 0.7@xmath35 .
figure 2(a ) shows the intensity map at @xmath8 measured with 135 ev photons , while panel ( b ) shows the calculated fs for @xmath36 .
the reasonable agreement in the overall shapes of the features indicates that we have made the correct assignment of @xmath23 .
the fermi surfaces shown in figs . 1 and 2 consist of several sheets .
to examine their character , we have plotted in fig . 3 ( upper panels ) the experimental band dispersions along the four different momentum cuts # 1 , # 2 , # 3 and # 4 shown in fig .
the band dispersion plot in cut # 1 is from ( @xmath37 , 0 ) to ( @xmath38 , 0 ) , which shows a small electron pocket at the center of the bz ( @xmath5 point ) with fermi momenta @xmath39 .
the bottom of this electron pocket is located about 300 mev below @xmath8 and it touches the top of a fully occupied band .
an appearance of the electron pocket at @xmath5 is expected based on a doping study of ba(fe@xmath40co@xmath32)@xmath0as@xmath0 ( ref . ) since co provides one extra electron and shifts @xmath8 upwards in energy .
in addition , two smaller pockets are observed at fermi momenta @xmath41 .
the band dispersion plot in cut # 2 is from @xmath42 to @xmath43 and shows larger electron pockets at the corner of the first bz at @xmath44 with @xmath45
. cut # 3 from @xmath46 to @xmath47 shows a large electron pocket that has fermi momenta similar to the previous one .
finally , we plot cut # 4 , from @xmath48 to @xmath49 , which again shows a small electron pocket at @xmath5 and large electron pockets at the @xmath6 points . in the lower panels , we show the corresponding calculated band dispersions , which are in reasonable qualitative agreement with the experimental data .
we also note the presence of a rather flat portions of the bands in close proximity to @xmath8 visible in both the data and calculations for cuts # 3 and # 4 .
these likely lead to an increase of the density of states at @xmath8 and therefore should have consequences for transport and thermodynamical properties . to further demonstrate this , we plot in fig . 4 ( left panel ) the energy distribution curves ( edcs ) in the range of momentum where this flat band is located ( marked by a white bar at the top of cut # 3 in fig .
the edc peaks are nearly dispersionless and located in very close proximity to @xmath8 .
the intensity of these peaks is very high , approaching the intensity of the valence band , as shown in the right panel of fig .
4 . in fig . 5
, we compare the intensity maps of baco@xmath0as@xmath0 measured at different binding energies .
it is seen that the electron pocket at the center of the bz shrinks to a point at about 400 mev below @xmath8 [ fig .
below this energy a hole - like contour is present that increases in size with increasing binding energy . at about 500 mev below @xmath8
the size of this hole contour is comparable to the hole fs sheet observed in bafe@xmath0as@xmath0 .
similarly , the electron pocket at the corner of the bz shrinks in size with increasing binding energy and reaches the size comparable to the feas - based material at about 500 mev binding energy .
in fact the intensity plot in fig .
5(f ) looks very much like the fs of unsubstituted bafe@xmath0as@xmath0 .
this indicates that substitution of co for fe mostly shifts the chemical potential without significantly changing the band dispersions .
thus , fully substituting the fe by co corresponds to an upward shift of the chemical potential by @xmath9500 mev . in fig . 6
, we compare the fermi surface of ba(fe@xmath40co@xmath32)@xmath0as@xmath0 ( adopted from ref . ) with constant energy intensity plots at high binding energy of baco@xmath0as@xmath0 samples .
the values of binding energies were chosen to match roughly the size of the contours of the electron bands in the upper and lower panels ( fig .
6 ) , respectively .
we note that the plots at high binding energy of baco@xmath0as@xmath0 quite well match the shape and size of the fermi surface of ba(fe@xmath40co@xmath32)@xmath0as@xmath0 , implying that a rigid band shift is a good approximation for low lying bands .
however , at lower binding energies in baco@xmath0as@xmath0 and higher cobalt concentrations in ba(fe@xmath40co@xmath32)@xmath0as@xmath0 , the shape and size of the contour at the zone center are quite different , despite the good match at the zone corners .
this signifies that the rigid band shift breaks down above -460 mev .
we presented an angle - resolved photoemission study of the fermi surface and band structure of baco@xmath0as@xmath0 and compared these with theoretical calculations carried out using the fp - lapw method .
reasonable qualitative agreement in the shape of the fs and band dispersion plots between the experiment and calculations was found .
the fermi surfaces consist of electron pockets both at the center and corner of the bz .
the size of these pockets is very different prohibiting interband nesting .
the measurements performed at higher binding energies ( @xmath9500 mev ) reveal intensity contours that are almost identical to the fermi surface of bafe@xmath0as@xmath0 .
this agreement indicates that full substitution of fe by co results in a chemical potential increase of about 500 mev without significant renormalization in the band dispersions .
we thank aaron bostwick and eli rotenberg for excellent support at the als and abhishek pandey for helpful discussions . research supported by the u.s .
department of energy , office of basic energy sciences , division of materials sciences and engineering .
ames laboratory is operated for the u.s .
department of energy by iowa state university under contract no .
de - ac02 - 07ch11358 .
the advanced light source is supported by the office of basic energy sciences , u. s. department of energy under contract no .
de - ac02 - 05ch11231 .
l. x. yang , y. zhang , h. w. ou , j. f. zhao , d. w. shen , b. zhou , j. wei , f. chen , m. xu , c. he , y. chen , z. d. wang , x. f. wang , t. wu , g. wu , x. h. chen , m. arita , k. shimada , m. taniguchi , z. y. lu , t. xiang , and d. l. feng , phys .
rev . lett . * 102 * , 107002 ( 2009 ) .
y. zhang , j. wei , h. w. ou , j. f. zhao , b. zhou , f. chen , m. xu , c. he , g. wu , h. chen , m. arita , k. shimada , h. namatame , m. taniguchi , x. h. chen , and d. l. feng , phys .
lett . * 102 * , 127003 ( 2009 ) .
t. kondo , r. m. fernandes , r. khasanov , chang liu , a. d. palczewski , ni ni , m. shi , a. bostwick , e. rotenberg , j. schmalian , s. l. budko , p. c. canfield , and a. kaminski , phys .
b * 81 * , 060507(r ) ( 2010 ) .
chang liu , t. kondo , r. m. fernandes , a. d. palczewski , e. d. mun , n. ni , a. n. thaler , a. bostwick , e.rotenberg , j. schmalian , s. l. budko , p. c. canfield and adam kaminski , nature phys . * 6 * , 419 ( 2010 ) .
chang liu , a. d. palczewski , r. s. dhaka , t. kondo , r. m. fernandes , e. d. mun , h. hodovanets , a. n. thaler , j. schmalian , s. l. budko , p. c. canfield , and a. kaminski , phys .
b * 84 * , 020509(r ) ( 2011 ) . | we use angle - resolved photoemission spectroscopy and full - potential linearized augmented - plane - wave ( fp - lapw ) calculations to study the electronic structure of baco@xmath0as@xmath0 . the fermi surface ( fs ) maps and the corresponding band dispersion data ( at 90 k and 200 k )
reveal a small electron pocket at the center and a large electron pocket at the corner of the brillouin zone .
therefore the nesting between electron and hole fs pockets is absent in this compound , in contrast to the parent compounds of feas - based high-_t_@xmath1 superconductors .
the electronic structure at about 500 mev binding energy is very similar to features at the chemical potential in bafe@xmath0as@xmath0 .
this indicates that complete substitution of co for fe causes a nearly rigid shift in the fermi energy by adding two electrons per formula unit without significant changes in the band dispersions .
the experimental fs topology as well as band dispersion data are in reasonable agreement with the fp - lapw calculations . |
fullerenes , hollow clusters made up of carbon atoms bonded by sp@xmath0 orbitals , have interesting conducting properties .
the origin of such a behavior is their delocalized @xmath1 frontier molecular orbitals , what gives rise to a high conductance when an electric field is applied on the molecule , e.g. by an external potential bias .
it has been discussed that free - electron and tight - binding ( tb ) models can capture the main features of the electronic transport through nearly spherical fullerenes @xcite .
the high conductivity of c@xmath2 has lead to speculate on the possibility of considering it as a conducting spherical shell @xcite .
particularly , several authors have studied the molecular junctions of the c@xmath2 fullerene under different types of connections , such as , a substrate and a stm tip @xcite , one - dimensional leads @xcite , carbon nanotubes @xcite , gold clusters @xcite or break junctions @xcite .
furthermore , the stability and strong hybridization of c@xmath2 with metallic surfaces make it also a feasible anchoring group with high conductance @xcite . in the search of similar suitable molecular junctions , other larger icosahedral fullerenes
c@xmath3 from the same family @xmath4 ( @xmath5 integer ) have also been shown to be stable @xcite , while the conductance of others , such as c@xmath6 or its complexes have also been explored @xcite .
on the other hand , by doping with boron and nitrogen , fullerene - based molecular junctions were found to have negative resistance @xcite .
a number of methods have been applied to the study of energetics and stability of buckyonions , namely , icosahedral fullerenes encapsulated by larger ones @xcite .
carbon nano - onions are interesting structures between fullerenes and multi - wall carbon nanotubes , having high thermal stability and chemical reactivity compared to cnts .
they also have the characteristic high contact area and affinity for noble metals , what make them interesting as anchoring groups for molecular electronics .
for instance , it has also been shown that onion - like nanoparticles can be used as electrochemical capacitors , also called supercapacitors , with high discharge rates of up to three order of magnitud higher than conventional supercapacitors @xcite .
the static polarizability , closely related to the response of the electronic charges to applied static fields , have been studied for onions formed by members of the icosahedral @xmath7 family , using both phenomenological models and first - principle methods @xcite , showing their capability to partially screen static external electric fields .
the conductance of onion - like structures functionalized with sulfide - terminated chains has been measured between a gold substrate and a gold stm tip @xcite .
the present work is aimed to study the electronic transmission of single - wall and multi - wall fullerenes weakly attached to metallic leads . in the next section
we discuss the tb model and the influence of the curvature and finite size of the layers on it , as well as the green function - based method for the calculation of the transmission . in section iii , we present our results for the dependence of the transmission function on the electron energy @xmath8 .
we study how @xmath8 is affected by the relative angular orientation of the shells , the number of intershell connections included in the tb model and the number of shells of the onions .
finally in section iv , we summarize our conclusions on the systems studied .
the hamiltonian of a @xmath9-atom fullerene @xmath10 is described in the tb approximation with one @xmath1 orbital per site @xmath11 where the summation runs on nearest neighbor atom pairs @xmath12 , and the operator @xmath13 ( @xmath14 ) creates ( annihilates ) an electron in the @xmath1 orbital centered at the atom @xmath15 ( @xmath16 ) .
the constant on - site energy has been taken as zero . a single parameter @xmath17 is used for the hopping integral between nearest neighbor atoms for a given fullerene . although no bond dimerization effect is included , it has been shown that its sole effect is to slightly break some degeneracies in the spectrum with no major qualitative effect @xcite . in order to take into account the effect of the curvature of the shell for the various fullerenes , we consider the hopping parameter @xmath17 to be a function of the mean radius @xmath18 of the ( nearly ) spherical shell of @xmath9 c atoms and the mean inter - atomic distance @xmath19 @xcite @xmath20.\ ] ] the value @xmath21 ev is a suitable hopping for graphene and is chosen to correctly reproduce the homo - lumo gap for c@xmath2 , @xmath22 ev , as obtained from dft calculations @xcite .
although the tb hamiltonian only depends on the topology of the molecule ( i.e. , on the atoms bonded ) , the molecular geometry affects the hopping integral . .[d
r t ] mean interatomic c - c distances , mean radii and hopping integral for the single - wall icosahedral fullerenes of the families @xmath23 and @xmath24 ( @xmath5 integer ) . [ cols="^,^,^,^,^,^,^,^,^,^",options="header " , ] table [ d r t ] shows that the mean radius of the shell is the main geometrical variable , with the mean inter - atomic distance being approximately constant for the various fullerenes c@xmath3 of the families @xmath25 and @xmath26 ( @xmath5 integer ) .
the inter - shell separations are approximately constant for successive fullerenes of each family ( @xmath27 , close to the inter - layer separation in bulk graphite , for the former , and @xmath28 for the latter ) .
we shall study single - wall fullerenes and bilayer and trilayer onions composed by two and three concentric shells of icosahedral symmetry ( @xmath29 ) , c@xmath30@c@xmath31 . in onions
, it has been shown that the resulting structures for each shell are much the same those ones of the isolated fullerenes .
therefore , we assume the geometry for each shell in the onion to be the same as that of isolated c@xmath3 , taken from yoshida data base . the most stable mutual orientation is the one that preserves the @xmath29 symmetry of the composite system @xcite .
furthermore , in the onion - like family c@xmath32 , it is reasonable to assume the strength of their intershell interactions as similar to those between the layers of graphite , due to the similarity in their interlayer separation .
figure [ bilayer]a shows a scheme of a part of two adjacent parallel layers in graphite separated a distance of 3.35 .
each layer is constituted by two triangular lattices whose sites are denoted as a and b. the subscript 1 or 2 indicates which layer each site belongs . in graphite , the distances and number of nearest neighbors are exactly determined by the relative position between the plane infinite layers of identical geometry . in two - wall buckyonions , nevertheless , the inner and outer fullerenes have finite sizes with different number of atoms , different geometrical structures due of their different number of hexagons ( e.g. , in the family 60@xmath33 , every member have 12 pentagons ) and eventually different orientations from each other ( even though both have @xmath29 symmetry ) , as shown in figures [ bilayer]b and [ bilayer]c .
therefore , although it is usually thought of intershell interactions in buckyonions as locally similar to interlayer interactions in bilayer graphene , a more thorough consideration of the intershell hopping is needed due to the differences mentioned above . in graphite
the hopping parameter @xmath34 describes the covalent bonding arising from @xmath35 hybridization in a layer .
inter - layer van der waals interactions are described by the parameters @xmath36 ev ( hopping energy between atoms a1 and a2 ) , @xmath37 ev ( between b1 and b2 atoms ) and @xmath38 ev ( hopping energy for a1-b2 and a2-b1 pairs ) @xcite . in bilayer onions ,
no such two lattices exists due to the finite size and the curvature of the layers . nevertheless , similar van de waals interactions between nearest neighbor ( nn ) and next - nearest neighbors ( nnn ) atoms exist . taking into account that for graphite @xmath39 and @xmath40 , we take a single value @xmath41 ev for the hopping between pairs of the nn and nnn between layers @xcite . the intershell interaction is included through the hopping integral @xmath42 between pairs of nn and nnn atoms as follows : we define a cutoff radius @xmath43 such that every pair of atoms , at @xmath44 and @xmath45 , belonging to adjacent shells and separated a distance shorter than @xmath43 is assigned a hopping @xmath42 , i.e. ,
@xmath46 it should be noted that , due to the faceting of the icosahedral symmetry , the number of intershell connections is not isotropic . as an illustration ,
figures [ bilayer]b and [ bilayer]c show two portions of a trilayer onion within a solid angle around the directions joining the centers of the pentagons and hexagons , respectively .
the lines joining atoms in adjacent shells represent the intershell connections @xmath47 for a given cutoff radius . due to the non sphericity of the shells ,
the outermost pentagon is not connected but the outermost hexagon it is .
we characterize the number of pairs of atoms connected at a given @xmath48 , by the mean number of neighbors ( in the outer shell ) ` felt ' by atoms in the inner shell , @xmath49 where @xmath50 is the number of intershell connections for a given @xmath48 , and @xmath51 is the number of atoms in the inner shell .
figure [ bilayer]d shows the variation of @xmath52 with @xmath43 for c@xmath2@c@xmath53 with @xmath29 symmetry , c@xmath53@c@xmath54 with @xmath29 symmetry , c@xmath2@c@xmath53 without symmetry ( symmetry @xmath55 ) and for two adjacent layers of graphite . in the latter ,
there are wide ranges of @xmath43 where the number of connections keeps constant .
thus , due to the parallel orientation of the infinite layers , there are only some discrete values at which the number of connections increases due to the inclusion of neighbors farther to a given atom .
for the onions with @xmath29 symmetry , there are common ranges of @xmath48 that show step - like behavior for both onions ( see regions shadowed in figure [ bilayer]d ) . in the next section we present results of calculations with a number of intershell connections ( characterized by @xmath52 ) chosen in those regions .
such a step - like dependence is in contrast to the approximately linear dependence for c@xmath2@c@xmath53 with symmetry @xmath55 , when the symmetry axes of the @xmath29 fullerenes are misaligned .
the faceting of the fullerenes induced by the @xmath29 symmetry is particularly noticeable for the largest layers , as can be seen en figure [ bilayer]e for c@xmath2 , c@xmath53 and c@xmath54 .
therefore , the fraction of carbon atoms having a nn or nnn within the range @xmath56 is smaller for the larger fullerenes .
hence , the tb hamiltonian for the onions becomes that of the isolated layers with an inter - layer interaction term @xmath57 where site @xmath15 belongs to a inner shell , and site @xmath16 is its nn or nnn on the adjacent outer shell .
when a molecule is attached between two metallic leads and subject to a potential bias , the charge current flowing through it can be calculated with the landauer equation @xcite @xmath58,\ ] ] where @xmath59 and @xmath60 are the fermi distributions at the left ( l ) and right ( r ) leads . at low temperatures
, the transmission function represents the dimensionless conductance ( in units of the quantum @xmath61 ) and is calculated as @xmath62 where @xmath63 and @xmath64 are the matrix representation of the advanced and retarded green functions @xmath65 , and @xmath66 and @xmath67 are the spectral densities of the leads @xcite . in the wide band approximation , the green function of the connected system can be obtained by using dyson equation , as @xmath68 where @xmath69 is the retarded green function of the isolated system , and @xmath70 are the self - energies of the leads , considered to be energy independent . throughout this work ,
we connect the fullerenes to the leads , using @xmath71 ev , through two carbon atoms located diametrally opposite to each other , one atom in the vertex of a pentagon and its corresponding one obtained by applying the operation of inversion with respect to the center of symmetry .
the approximately spherical form of the fullerenes , particularly the smaller ones , allows to treat onions as a family of concentric spherical shells for the calculation of some properties , such as the determination of radii of equilibrium , intershell distances , static polarizability or photoionization cross section @xcite . for other applications
, however , a more accurate description of their geometry is relevant . in particular , the trend in larger fullerenes to approach faceted icosahedral forms makes very relevant the relative orientation of adjacent layers , even for concentric shells having individually @xmath29 symmetry , as already shown in figure [ bilayer]d for the average number of connections per atom @xmath52 .
the influence of the relative orientation on the quantum transmission is shown in figure [ orientation ] for two bilayer onions : c@xmath2@c@xmath53 , and c@xmath53@c@xmath54 .
two orientations were considered : one where the onion has overall @xmath29 symmetry , that is , with the symmetry axes of the individual shell aligned ; and one where both shells are concentric but the individual symmetry axes are rotated an arbitrary relative angle from each other ( @xmath55 symmetry ) . the differences in @xmath8 between the icosahedral ( @xmath29 ) and the non symmetrical ( @xmath55 ) onion are visible in figure [ orientation ] .
in the onion with misalignment between the shells , some degeneracies in the energy spectrum are broken , as reflected in the occurrence of multiple resonant states with many peaks and antiresonances close to each other leading to a trend to the formation of narrow bands for the larger onion c@xmath53@c@xmath54 .
the transmission function of the non symmetrical onions have rapid variations from perfect to vanishing transmission with slight variations of the fermi energy @xmath72 , while @xmath8 for the @xmath29 onions is a well behaved smooth function with a few peaks of perfect transmission in the range around the homo - lumo gap .
the gap itself increases when the shells become disoriented from each other what also corresponds to a less stable configuration , as previously reported @xcite . as mentioned in section [ two - shells ] , the relative orientation of hexagons belonging to adjacent shells , the choice of the cutoff radius for defining nn and nnn sites of a given atom and the faceting of the larger fullerenes preclude a unique definition of the number of intershell connections to be included in the hamiltonian ( [ h_onion ] ) .
both the interlayer distance in graphite and intershell distance in the @xmath73 family are close to 3.5 .
thus , we took two cutoff radii close to this value and a larger one ( see shadowed regions of figure [ bilayer]d ) , to include a bigger number of nn and nnn pairs in the intershell interaction hamiltonian . in the following
we show the results of calculated @xmath8 for icosahedral onions ( @xmath29@@xmath29 ) and the aforementioned three choices of @xmath74 . in those three regions , @xmath75 , 1 and 4 for c@xmath2@c@xmath53 , and @xmath76 , 1 and 2 for c@xmath53@c@xmath54 . in figure
[ r_cutoff ] , the transmission @xmath8 is depicted for the three values of @xmath52 , as indicated in the legends . it should be noticed that for the smaller onion c@xmath2@c@xmath53 , the three @xmath52 values give almost the same curve ( fig .
[ r_cutoff]a ) .
therefore , this increase of connectivity does not affect the transmission throughout the composite system .
therefore , the main paths of transmission between shells are along the pairs formed by the atoms of the inner shell with its closest neighbor in the outer one .
figure [ r_cutoff]b shows that the dependence of the transmission on the number of intershell connections is more important for the larger onion c@xmath53@@xmath77 . for larger fullerenes the deviation from the spherical shape becomes more noticeable as the number of atoms increases .
thus the c@xmath53 shell is less spherical than c@xmath2 , and c@xmath54 is clearly faceted .
this departure from sphericity favours the hopping from the inner to the outer shell along certain directions , namely , those in which the icosahedral faces of both shells are closer to each other .
therefore , the larger the number of intershell links the better the quantum transmission .
interestingly , the increasing in the number of connections mainly affect the states above the highest occupied - lowest unoccupied ( homo - lumo ) gap , noticeably the lumo and lumo+1 ones . finally , figure [ r_cutoff]c shows @xmath8 for the trilayered onion c@xmath2@c@xmath53@@xmath77 which is notoriously similar to the one of the bilayer onion of fig [ r_cutoff]b , thus showing that the two external shells are the most relevant for the transmission , with only small corrections coming from the innermost c@xmath2 shell .
this effect is discussed in greater detail in section [ shells ] .
the bilayer onion have four peaks above the gap in range shown , namely , those corresponding to lumo , lumo+1 , lumo+2 and lumo+3 .
the two central peaks ( lumo+1 , lumo+2 ) are almost degenerates in c@xmath53@@xmath77 but become better resolved in the trilayered onion .
we shall show here that the outermost shell greatly determines the most relevant features of the transmission spectra of bilayered and trilayered onions .
the importance of the influence of the inner shells on @xmath8 decreases inwards . in figure
[ r_cutoff]c we observed that the most noticeable effects when adding c@xmath2 to the two external shells are the occurrence of an antiresonance between the lumo and lumo+1 peaks , and the widening of the narrow fano - like profile after the lumo+2 peak ( at @xmath78 ev . ) .
such a type of effects are analogous to the modifications to the transmission function through a chain introduced by adding a lateral site or chain to the system . in the following we fix the number of connections in trilayer onions by chosing @xmath79 ( @xmath80 ) between the two outermost ( innermost ) shells .
figure [ shell number ] shows the transmission for the bilayer onions c@xmath2@c@xmath53 and c@xmath53@c@xmath54 as compared to those for the corresponding single - wall fullerenes .
it can be seen that for energies below the gap , @xmath8 for the onions is very similar to the one of the outer fullerene , i.e. , the one for c@xmath53 in figure [ shell number](a ) , and for c@xmath54 in figure [ shell number](b ) . the peaks above the gap preserves similar features both in c@xmath2@c@xmath53 and c@xmath53 , although with a relative energy shift of the peaks . in figure [ shell number]b similar considerations can be made about the comparison the transmission through c@xmath53@c@xmath54 and c@xmath54 ; the latter provides most of the features observed in the former , particularly for energies below the gap .
the intershell connections in c@xmath53@c@xmath54 eventually contributes to the occurrence of antiresonances , such as that at @xmath81 ev , not present in @xmath8 for c@xmath54 . in other cases
, it softens the vanishing of transmission , such as in the antiresonance of c@xmath54 at @xmath82 ev which becomes in a finite transmission for c@xmath53@c@xmath54 . roughly speaking , introducing c@xmath2 as a third innermost shell
does not greatly modify the transmission of the bilayered onion .
the influence of the external shell on the transmission can be interestingly shown in the onions formed from fullerenes belonging to the family @xmath83 , which also have icosahedral symmetry .
thus , c@xmath3 fullerenes with @xmath84 , 80 , 180 , 320 , @xmath85 are predicted to be stable , with a radii increasing by approximately 2 from each member from the family to the next one .
the radii of equilibrium for the family @xmath73 were shown to be accurately determined by a continuous spherically symmetric lennard - jones model @xcite .
application of the same model for the multiwall onions of the @xmath83 family ( with @xmath5 odd integer ) , results an intershell distance of about 4 , not far from the intershell distance for the @xmath73 family or the interlayer distance in graphite .
therefore , we show calculations for the single - wall , two - wall and three - wall fullerenes obtained from c@xmath6 , c@xmath86 and c@xmath87 .
our tb calculations show that c@xmath6 and c@xmath87 are gapless , while c@xmath86 presents a gap of @xmath88 ev . in figures
[ family 20]a and [ family 20]b , c@xmath86 is the outer and inner shell respectively . as a consequence , in figure [ family 20]a
, c@xmath6@c@xmath86 shows a region of vanishing transmission within the gap , while it keeps finite for c@xmath86@c@xmath87 , except for the already discussed narrow antiresonances arising from the intershell connections .
hence , the transmission function is strongly sensitive to the electronic structure of the external shell . as a final example of this property , figure
[ family 20]b , shows the transmission for the trilayered c@xmath6@c@xmath86@ c@xmath87 as compared to those of the single - wall fullerenes .
it is seen that the most of the features of the onion are reproduced by the transmission through c@xmath87 , with some fano - like antiresonances originated in the transmission spectrum of the inner c@xmath86 and , to less extent , in the innermost c@xmath6 .
it can bee seen that the three - wall onion ( figure figure [ family 20]b ) is well described by the two - wall one ( figure [ family 20]a ) .
interestingly , the fano - like resonance of c@xmath6@c@xmath86@ c@xmath87 at @xmath89 ev is not present in c@xmath86@ c@xmath87 but it is in c@xmath6@c@xmath86 ; therefore , such a peak of conductance is an effect from the coupling of the two inner sheells .
in this work we have studied theoretically the quantum transmission through single - wall fullerenes and bilayered and trilayered onions of icosahedral symmetry , when attached to metallic leads , by using a tb hamiltonian and green functions methods .
although the van der waals interactions between onion shells are supposed to be similar to those between graphite layers , the finite size of the fullerenes , their curvature and relative orientations need some analysis for including the intershell hopping parameter .
we include in the model the effect of finite size and curvature through a parametrization of the hopping integral as a function of the number of atoms of the shell .
the number of connections from a given atom to others belonging to adjacent shells was studied by introducing a cutoff radius for the interaction .
we found that misalignment of the symmetry axes produces breaking of the level degeneracies of the individual shells , giving rise some narrow quasi - continuum bands instead of the localized discrete peaks of the individual fullerenes .
most of the features of the transmission through the onions are already visible in the transmission function of the single - wall fullerene forming the outer shell .
the main modifications between them are antiresonances arising from the coupling between the outer layer with the next innermost one . for three - wall onions ,
the transmission becomes barely sensitive to the most internal shell .
interestingly , when the fullerene of the external shell is gapless , the transmission of the onion does not vanish along finite ranges of energy .
this property could be useful for designing multilayered fullerenes with tailored conductance by properly growing the outermost layers .
we acknowledge financial support for this project from conicet ( pip 11220090100654/2010 ) and sgcyt(unne ) through grant pi f007/11 .
99 e. manousakis , phys .
b * 44 * , 10991 ( 1991 ) .
n. mizorogi , m. kiuchi , k. tanaka , r. sekine , and k. aihara , chem .
. lett . * 378 * , 598 ( 2003 ) .
m ya . amusia and a. s. baltenkov , phys .
lett . a * 360 * , 294 ( 2006 ) .
m. paulsson and s. stafstrm , j. phys .
matter * 11 * , 3555 ( 1999 ) .
n. nel , j. krger , l. limot and r. berndt , nano lett .
* 8 * , 1291 ( 2008 ) .
a. saffarzadeh , j. appl . phys . * 103 * , 083705 ( 2008 ) . a. a. shokri and sh .
nikzad , j. appl .
* 110 * , 024303 ( 2011 ) .
s. bilan , l. a. zotti , f. pauly and j. c. cuevas , phys . rev .
b * 85 * , 205403 ( 2012 ) .
e. lrtscher , v. geskin , b. gotsmann , j. fock , j. k. srensen , t. bjrnholm , j. cornil , h. s. j. van der zant and h. riel , small * 9 * , 209 ( 2013 ) . m. yu , i. chaudhuri , c. leahy , s. y. wu , and c. s. jayanthi , j. chem
. phys . *
130 * , 184708 ( 2009 ) .
b. i. dunlap and r. r. zope , chem .
. lett . * 422 * , 451 ( 2006 ) .
m. otani , t. ono and k. hirose , phys .
b * 69 * , 121408(r ) ( 2004 ) . y .-
an , ch .- l .
yang , m .- s .
wang , x .-
ma and d .- h .
wang , j. chem .
phys . * 131 * , 024311 ( 2009 ) .
g. ji , y. xu , b. cui , ch .
fang , x. kong , d. li and d. liu , rcs advances * 2 * , 11349 ( 2012 ) . m. yaghobi and h. r. vanaie , phys .
a * 375 * , 1249 ( 2011 ) .
a. maiti , c. j. brabec and j. bernholc , phys .
* 70 * , 3023 ( 1993 ) .
h. gurin , j. phys .
b * 30 * , l481 ( 1997 ) .
m. i. heggie , m. terrones , b. r. eggen , g. jungnickel , r. jones , c. d. latham , p. r. briddon , h. terrones ( 1997 ) .
pennington , nj : the electrochemical society recent advances in the chemistry and physics of fullerenes and related materials , montreal , canada : fullerenes : chemistry , physics , and new directions ix 4 , pp .
1141 - 1150 .
m. i. heggie , m. terrones , and b. r. eggen , g. jungnickel , r. jones and c. d. latham , p. r. briddon , h. terrones , phys .
b * 57 * , 13339 ( 1998 ) .
h. dodziuk , g. dolgonos , o. lukin , chem phys . lett . * 329 * , 351 ( 2000 )
. o. e. glukhova , a. i. zhbanov and a. g. rezkov , phys .
solid state * 47 * , 390 ( 2005 ) .
d. baowan , n. thamwattana and j. m. hill , eur .
j. d * 44 * , 117 ( 2007 ) .
a. n. enyashin and a. l. ivanovskii , phys . solid state * 49 * , 392 ( 2007 ) .
m. pudlak and r. pincak , phys .
a * 79 * , 033202 ( 2009 ) .
o. p. charkin , russ .
chem . * 58 * , 46 ( 2013 ) .
bing - she xu , new carbon materials * 23 * , 289 ( 2008 ) and refs . there in . d. pech , m. brunet , h. durou , p. huang , v. mochalin , y. gogotsi , p .-
taberna and p. simon , nature nanotechnology * 5 * , 651 ( 2010 ) .
s. iglesias - groth , a. ruiz , j. bretn and j. m. gomez llorente , j. chem .
phys . * 118 * , 7103 ( 2003 ) .
r. zope , j. phys .
b * 41 * , 085101 ( 2008 ) .
g. k. gueorguiev , j. m. pacheco and d. tomnek , phys .
* 92 * , 215501 ( 2004 ) . s. sek , j. breczko , m. e. plonska - brzezinska , a. z. wilczewska and l. echegoyen , chem .
* 14 * , 96 ( 2013 ) .
dresselhaus et al .
_ science and application of nanotubes _ , edited by d. tomarek and r. j. endbody ( kluwer / academic / dordrecht / new york , 2002 ) . s. saito and a.
oshiyama , phys .
* 66 * , 2637 ( 1991 ) .
a. h. castro neto , f. guinea , n. m. r. peres , k. s. novoselov and a. k. geim , rev .
phys . * 81 * , 109 ( 2009 ) .
e. v. castro , k. s. novoselov , s. v. morozov , n. m. r. peres , j. m. b. lopes dos santos , j. nilsson , f. guinea , a. k. geim and a. h. castro neto , j. phys .
matter * 22 * , 175503 ( 2010 ) .
r. landauer , phys .
a * 85 * , 91 ( 1986 ) ; m. btikker , phys .
* 57 * , 1761 ( 1986 ) .
j. c. cuevas and e. scheer , molecular electronics , world scientific , singapore ( 2010 ) .
y. b. xu , m. q. tan and u. becker , phys .
* 76 * , 3538 ( 1996 ) .
a. ruiz , j. bretn , and j. m. gomez llorente , j. chem .
phys . * 120 * , 6163 ( 2004 ) .
o. kidun , n. fominykh , and j. berakdar , comp . mat .
35 * , 354 ( 2006 ) .
v. k. dolmatov , p. brewer and s. t. manson , phys .
a * 78 * , 013415 ( 2008 ) . | we use a tight - binding hamiltonian and green functions methods to calculate the quantum transmission through single - wall fullerenes and bilayered and trilayered onions of icosahedral symmetry attached to metallic leads .
the electronic structure of the onion - like fullerenes takes into account the curvature and finite size of the fullerenes layers as well as the strength of the intershell interactions depending on to the number of interacting atom pairs belonging to adjacent shells .
misalignment of the symmetry axes of the concentric iscosahedral shells produces breaking of the level degeneracies of the individual shells , giving rise some narrow quasi - continuum bands instead of the localized discrete peaks of the individual fullerenes . as a result ,
the transmission function for non symmetrical onions are rapidly varying functions of the fermi energy .
furthermore , we found that most of the features of the transmission through the onions are due to the electronic structure of the outer shell with additional fano - like antiresonances arising from coupling with or between the inner shells . |
x ray observations are excellent probes of coronally active and accreting sources . whereas studies of x ray sources in our galaxy have to a large extent focussed on bright systems , investigations of fainter source classes
have typically been done in the galactic center ( e.g. @xcite ) . there , however , crowding and extinction make it more difficult to identify the correct optical and/or infrared counterparts for large fractions of the sources ( e.g. @xcite ) .
it is clear from previous surveys that multi wavelength observations are vital for classifying faint x
ray sources since x ray spectral information alone is rarely sufficient .
classification is important for various science goals as we outline below .
the galactic bulge survey ( gbs ) we present in this paper was designed to allow multi
wavelength observations of the detected x ray sources .
the gbs consists of and optical imaging of two strips of @xmath5 , one centered @xmath6 above the galactic plane and the other @xmath6 below the plane .
we have chosen this area as the source density is still high , but by excluding @xmath7 we avoid the regions that are most heavily affected by extinction and source confusion ( see figure [ changbs ] ) . in table
[ tab : reference ] we provide a reference table for the acronyms used in this paper .
.reference table for the acronyms used in this paper .
[ cols= " < , < " , ] @xmath8 multiple sources from the rosat pointed source catalog are consistent with the position .
we list the nearest one .
+ @xmath9 these two sources are 4.8 apart ; these are likely two detections of the same source .
we have started the galactic bulge survey ( gbs ) with the goals : i ) to identify quiescent , eclipsing ns and bh x ray binaries that are bright enough in the optical or near infrared to allow phase resolved spectroscopic observations tailored to measure the compact object mass .
ii ) to constrain binary population synthesis and x ray binary formation and evolution models by means of a source number count and a study of the spatial distribution of x ray binaries . in this paper
we have presented the source list and some properties of the x ray sources of observations covering @xmath10two - thirds ( @xmath11 square degrees ) of the total envisaged survey area of 12 square degrees .
the accurate source position will help identify the optical counterparts .
the 1234 x ray sources that have been discovered so far compares well with the total number of @xmath12 x ray sources that we predict we should detect in the full 12 square degrees . however , this is of course no guarantee that the number of sources per source class are close to those we calculated .
we compared our source list with the source list of the rass .
two bsc rass sources , 1rxs j175113.2 - 293842 and 1rxs j174220.8 - 273736 , are not detected in our gbs survey , indicating a decrease in flux with factors larger than 200 and 500 , respectively .
furthermore , we compared our source list with the sources found in the catalog of sources derived from pointed hri and pspc rosat observations that fall inside the gbs area ( see table 3 ) .
we also imaged the complete survey area using optical observations obtained with the 4 m blanco telescope at ctio .
furthermore , we re - imaged the survey area @xmath1330 times in the @xmath14 band to search for ( periodic ) variable sources .
results of these campaigns will be presented in forthcoming papers .
we are in the process of obtaining optical spectroscopic and photometric observations of all optical proposed counterparts in the error circle to classify the x ray sources .
this step is crucial to achieve our science goals outlined above .
a full discussion of all known information on all the x ray sources is beyond the scope of this paper and we defer the discussion of the other fainter sources to a forthcoming paper , where they will be discussed together with optical photometric and spectroscopic information .
p.g.j . and g.n .
acknowledge support from the netherlands organisation for scientific research . r.i.h . ,
c.t.b , v.j.m . , and l.g .
would like to acknowledge support from national science foundation grant no .
ast-0908789 and nasa / louisiana board of regents grant nnx07at62a / leqsf(2007 - 10 ) phase3 - 02 .
coh is supported by nserc .
ds acknowledges an stfc advanced fellowship .
tjm and ad thank stfc for support via a rolling grant to the university of southampton . | we introduce the galactic bulge survey ( gbs ) and we provide the source list for the region that has been observed to date . among the goals of the gbs are constraining the neutron star equation of state and the black hole mass distribution via the identification of eclipsing neutron star and black hole low mass x ray binaries . the latter goal
will , in addition , be obtained by significantly enlarging the number of black hole systems for which a black hole mass can be derived .
further goals include constraining x ray binary formation scenarios , in particular the common envelope phase and the occurrence of kicks , via source - type number counts and an investigation of the spatial distribution of x ray binaries , respectively
. the gbs targets two strips of @xmath0 ( 12 square degrees in total ) , one above ( @xmath1 ) and one below ( @xmath2 ) the galactic plane in the direction of the galactic center at both x ray and optical wavelengths . by avoiding the galactic plane ( @xmath3 ) we limit the influence of extinction on the x ray and optical emission but still sample relatively large number densities of sources .
the survey is designed such that a large fraction of the x ray sources can be identified from their optical spectra .
the x ray survey , by design , covers a large area on the sky while the depth is shallow using 2 ks per pointing . in this way
we maximize the predicted number ratio of ( quiescent ) low mass x ray binaries to cataclysmic variables .
the survey is approximately homogeneous in depth to an 0.5 - 10 kev flux of 7.7@xmath4 .
so far , we have covered about two thirds ( 8.3 square degrees ) of the projected survey area with providing over 1200 unique x
ray sources .
we discuss the characteristics and the variability of the brightest of these sources . |
observations motivate the notion of `` cosmic downsizing '' ( as coined by * ? ? ?
* ) , with the global star formation rate declining rapidly below @xmath4 , and the sites of galactic star formation shifting to smaller masses at lower redshift .
moreover , galaxy surveys such as sdss , combo-17 , and deep2 demonstrate that the color distribution of galaxies is bimodal ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , and that this bimodality extends at least to @xmath5 ( e.g. , * ? ? ?
* ; * ? ? ?
it is increasingly established that high mass , red elliptical galaxies have older stellar populations than smaller spheroids ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
but , many studies also see a significant population of massive / luminous galaxies in place ( i.e. assembled ) by @xmath4 ( e.g. , * ? ? ?
* ; * ? ? ?
* and references therein ) , with measurements of galaxy stellar mass functions ( mfs ) and luminosity functions at redshifts @xmath6 favoring either a uniform increase or buildup in the numbers of early - type ( `` red sequence , '' rs ) galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ) or an anti - hierarchical scenario in which this `` buildup '' at @xmath7 occurs primarily at the low - mass end of the rs @xcite .
the blue , disk dominated , star forming galaxy mass function ( dominant at low mass ) , meanwhile , remains relatively constant , or perhaps declines to @xmath8 . as a consequence , the `` transition mass , '' above which the red galaxy population dominates the galaxy mf , decreases with time , tracing this downsizing trend .
there is evidence for some evolution at the highest masses as ellipticals grow by spheroid - spheroid or `` dry '' mergers @xcite , but this , by definition , proceeds strictly hierarchically , and _ can not
_ account for the movement of mass onto the rs in the first place or any buildup in the number density of low - mass ellipticals .
meanwhile , the discovery of tight correlations between the masses of central supermassive black holes ( bhs ) in galaxies and the bulge or spheroid stellar mass @xcite , velocity dispersion @xcite or concentration @xcite implies that the formation of galaxies and bhs must be linked .
moreover , the evolution of the quasar luminosity function ( qlf ) shows a sharp decline after @xmath4 , with the density of lower - luminosity agn peaking at low redshift ( e.g. , * ? ? ? * and references therein ) . to the extent that bh
assembly traces galaxy assembly ( i.e. there is weak evolution in the bh - host mass relation , as observed to at least @xmath9 by e.g. @xcite ) , this implies early _ assembly _ times ( @xmath9 ) for many of the most massive systems containing @xmath10 bhs .
a number of theoretical models have been proposed to explain the evolution of these populations with redshift , and their correlations with one another ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in many of these models , the merger hypothesis @xcite provides a potential physical mechanism linking galaxy star formation , morphology , and black hole evolution and explaining these various manifestations of cosmic `` downsizing . '' in this scenario , gas - rich galaxy mergers channel large amounts of gas to galaxy centers ( e.g. , * ? ? ? * ; * ? ? ? * ) , fueling powerful starbursts ( e.g. , * ? ? ? * ; * ? ? ? * ) and buried bh growth ( e.g. , * ? ? ? * ; * ? ? ?
* ) until the bh grows large enough that feedback from accretion rapidly unbinds and heats the surrounding gas @xcite , leaving an elliptical galaxy satisfying observed correlations between bh and spheroid mass .
major mergers rapidly and efficiently exhaust the cold gas reservoirs of the progenitor systems , allowing the remnant to rapidly redden with a low specific star formation rate , with the process potentially accelerated by the expulsion of remnant gas by the quasar ( e.g. , * ? ? ?
this naturally explains the observed close association between the elliptical and red galaxy populations ( e.g. , * ? ? ?
in a qualitative sense , the evolution of the characteristic mass at which these processes occur can be understood as follows .
mergers proceed efficiently at high redshift , occurring most rapidly in the regions of highest overdensity corresponding to the most massive galaxies , building up the high - mass elliptical mf .
however , once formed these galaxies are `` dead '' , and mergers involving gas - rich galaxies must transition to lower masses
. recent hydrodynamical simulations , incorporating star formation , supernova feedback , and bh growth and feedback @xcite make it possible to study these processes self - consistently and have lent support to this general picture .
mergers with bh feedback yield remnants resembling observed ellipticals in their correlations with bh properties @xcite , scaling relations @xcite , colors @xcite , and morphological and kinematic properties @xcite .
the quasar activity excited through such mergers can account for the qlf and a wide range of quasar properties at a number of frequencies @xcite , and with such a detailed model to `` map '' between merger , quasar , and remnant galaxy populations it is possible to show that the buildup and statistics of the quasar and red galaxy populations are consistent and can be used to predict one another @xcite .
however , it is by no means clear whether this is , in fact , the dominant mechanism in the buildup of early - type populations and quasars and their evolution with redshift .
for example , many semi - analytic models incorporate quasar triggering / feedback and morphological transformation by mergers @xcite . however , some models tie quasar activity directly to star formation ( e.g. , * ? ? ?
* ) , implying it will evolve in a manner tracing star - forming galaxies , with this evolution and the corresponding downsizing effect roughly independent of mergers and morphological galaxy segregation at redshifts @xmath11 .
others invoke post - starburst agn feedback to suppress star formation on long timescales and at relatively low accretion rates through e.g. `` radio - mode '' feedback @xcite , which , if this is also associated with optical qso modes , would imply quasars should trace the established `` old '' red galaxy population at each redshift .
there are , of course , other sources of feedback , with galactic superwinds from star formation presenting an alternative means to suppress subsequent star formation , although the required wind energetics are sufficiently high to prefer a quasar - driven origin ( e.g. , * ? ? ?
several models invoke a distinction between `` hot '' and `` cold '' accretion modes @xcite , in which new gas can not cool into a galactic disk above a critical dark matter halo mass , potentially supplemented by agn feedback @xcite , as the dominant distinction between the blue cloud and red sequence , essentially independent of effects on scales within galaxies .
it is also important to distinguish the processes which may be associated with the initial movement of galaxies onto the red sequence from their subsequent evolution .
once morphologically transformed by a gas - rich merger , for example , mass can be moved `` up '' the rs ( galaxies increased in mass ) by gas - poor mergers , but it can not be _ added _ to the red sequence in this manner .
it also remains an important cosmological question to understand how , once formed , further growth of ellipticals by accretion or `` cooling flows '' may be halted .
the models above invoke various feedback processes , including `` radio mode '' activity ( e.g. , * ? ? ?
* ) , cyclic quasar or starburst - driven feedback @xcite , massive entropy injection from a single quasar epoch ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , and `` hot mode '' accretion @xcite to address this problem .
although critical to our understanding of galaxy formation , these processes must operate over timescales of order the hubble time for all massive galaxies once formed , and therefore are _ not _ necessarily associated with the _ addition _ of mass to the red sequence .
as such , the details of these long - term suppression mechanisms should be studied in different ( e.g. already formed elliptical ) populations , and are outside the scope of this paper .
observationally , it is still unclear whether mergers can account for the the buildup of elliptical and/or quasar populations ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) .
even within the context of the merger hypothesis , the relative importance of dissipational ( gas rich , disk ) vs. dissipationless ( gas poor , spheroid - spheroid ) mergers is unclear ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , although all measurements agree that the `` dry '' merger rate is much less than the gas - rich merger rate at all observed redshifts @xcite .
this is essentially related to the critical question of whether the buildup of the red sequence and elliptical populations is dominated by the formation or movement of `` new '' early - type galaxies onto that sequence or instead by the hierarchical assembly of small `` seed '' early - types and substructure formed at high redshift ( which will also not trigger quasar activity ) .
fundamentally , it is not clear and has not yet been tested whether the observed downsizing trends in the transition mass , galaxy stellar populations , quasars , and other populations are in fact _
quantitatively _ the same trend , or merely _ qualitatively _ similar .
this represents a key test which can distinguish between several of the various scenarios above . attempting to predict the values of this transition mass in an a priori cosmological manner
is inherently model dependent and , at least at low redshift , degenerate between the various models described above .
however , if mergers are indeed the critical link in the process causing the flow of galaxies from the blue to red sequence and triggering quasar activity , then it is a strong prediction of these theories , and specifically the modeling of hopkins et al .
( 2005a - d,2006a - d ) that the same mergers are responsible for the bulk of the bright quasar population and the buildup of the new mass on the red sequence at each redshift .
in other words , these downsizing trends must quantitatively reflect one another .
in this picture , the `` transition mass '' ( @xmath12 ) may represent the `` smoking gun '' of mergers causing the flow of galaxies from the blue to red sequence .
therefore , to the extent that @xmath12 traces the mass at which the red sequence is being `` built '' at some @xmath13 , it should also trace the characteristic mass of star - forming galaxies merging at that time , and the characteristic mass of galaxies hosting quasars which are initially triggered by those mergers .
of particular interest , the empirical test of this association does not require the adoption of some a priori model for galaxy formation . here , we consider the observed @xmath12 over the interval @xmath6 , and compare it to the characteristic masses of quasar hosts and merging galaxies over the same range in redshift .
we demonstrate that they appear to be evolving in a manner consistent with a merger - driven unification model of quasars , interacting galaxies , and the red galaxy population .
note that we use the term `` quasar '' somewhat loosely , as a proxy for high - eddington ratio accretion inevitably caused by gas - rich mergers , although there may be other triggering mechanisms as well @xcite .
such activity will of course be significantly weaker in small systems ( especially those typical of local ulirgs ) and may not technically qualify as a `` classical '' optical quasar @xcite , but this distinction is essentially arbitrary and has little impact on our analysis ( see also * ? ? ?
we adopt a @xmath14 , @xmath15 , @xmath16 cosmology .
all stellar masses are rescaled to a @xcite imf .
various studies have used different definitions and terms for the mass which separates the dominance of old , red , low - sfr elliptical galaxies from that of young , blue , star - forming disk galaxies .
it is also possible to divide the galaxy population along any one of those quantities .
although it has been established in a number of observational studies that the galaxy population is bimodal with respect to color , specific star formation rate , and morphology ( e.g. , * ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and that this bimodality extends at least to @xmath5 ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , it is still possible that the various definitions used to separate these bimodal distributions could result in a systematically different `` separation point . '' in what follows , we consider several definitions of the `` transition '' galaxy stellar mass in terms of the mfs ( @xmath17 ) of early and late type systems : the @xcite transition mass @xmath12 at which the density of early and late type systems are equal , @xmath18 the @xcite `` quenching '' mass @xmath19 at which the contribution of late types to the total mass function cuts off , @xmath20 and the @xcite `` downsizing '' mass @xmath21 above which @xmath22 of the @xmath8 rs mf has been assembled by a given redshift , @xmath23 for each , we consider a division between early and late types defined by either a color ( i.e. separating galaxies on the redshift - dependent rs from the `` blue cloud '' ; see e.g. @xcite ) , star formation rate ( generally from detailed population synthesis modeling , although we include looser emission / absorption galaxy spectral type separations ) , or morphology ( either by visual morphological identification or concentration / surface brightness ) criteria . above these masses , early type galaxies in the red , low - sfr , high - concentration , morphologically elliptical `` half '' of the bimodal distribution dominate the total galaxy mf , and below , late type galaxies in the blue , high - sfr , low - concentration , morphologically disk - like or irregular bimodal `` half '' dominate ( generally true also for @xmath21 , though it is technically independent of @xmath24 ) .
we determine these masses from a number of compiled type - separated mass functions , shown in figure [ fig : oplot.mfs ] ( for clarity , just the early - type mfs are shown ) .
data in all cases are converted to our adopted cosmology and rescaled to a @xcite imf .
at all redshifts , only points above the quoted completeness limits of each study are shown .
errors are as published , and generally account for cosmic variance similar to e.g. @xcite . at @xmath8 , we generally adopt the local mf determinations from @xcite from 2mass+sdss observations , but find no change in our results considering e.g. the 2mass+2dfgrs @xcite determination . at higher redshifts ,
our compilation includes ( * ? ? ?
* k20 ; @xmath25 ) , ( * ? ? ?
* k20 ; @xmath26 ) , ( * ? ? ?
* goods ; @xmath27 ) and ( * ? ? ?
* deep2 ; @xmath27 ) , ( * ? ? ?
* combo-17 ; @xmath28 ) , ( * ? ? ?
* cdf - s ; @xmath27 ) , and ( * ? ? ?
* fors deep+goods - s ; @xmath29 ) .
each of these considers the separate mfs of early and late types , divided according to at least one of the criteria above .
although many of these surveys cover a small area and thus may be subject to significant cosmic variance , our conclusions are unchanged if we restrict ourselves to the largest fields , and are in fact most robust for the wide - field surveys from combo-17 ( @xmath30 ) and deep2 ( @xmath31 ) .
figure [ fig : compare.methods ] shows resulting `` transition '' mass from these studies as a function of redshift , from the different definitions and galaxy type segregation methods above .
these samples generally derive masses from optical+near - ir spectral and photometric fitting . to compare
, we also consider the evolution in @xmath21 ( color - selected ) determined by @xcite .
they compile the @xmath32-band early - type ( red ) galaxy luminosity functions from combo-17 @xcite , deep2 @xcite , and the subaru/_xmm - newton _ deep survey @xcite at @xmath33 ( typical @xmath34 between luminosity functions ) , and use the redshift - dependent evolution of @xmath32-band mass - to - light ratios determined from fundamental plane studies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* roughly similar to mean formation redshifts @xmath35 ) to correct these to a @xmath8 equivalent @xmath32 luminosity and mass ( given the @xmath8 mass - dependent @xmath32/@xmath36-band @xmath37 ratios from @xcite ) .
we fit a simple relation of the form @xmath38 in each panel ( for illustrative purposes only , we do not intend for this to be considered a rigorous estimate of the evolution in these characteristic masses ) . despite the different selection and type separation methods and definitions of a characteristic mass , a nearly identical trend with redshift is recovered in every case .
in fact , the best - fit slopes @xmath39 for most selection methods and definitions are statistically indistinguishable from the cumulative best - fit slope , and several of the definitions agree nearly exactly in figure [ fig : compare.methods ] .
it is also reassuring that the indirect estimates from optical luminosity functions , which generally involve the largest samples and most finely probe the redshift evolution of @xmath21 , agree well at all redshifts with the mf estimates .
this suggests that the trend with redshift is real , and that it is independent of the potential systematics in sample selection , as these systematics can be quite different for the various criteria shown .
the absolute normalization of the `` transition '' mass does depend systematically on the definition chosen .
for example , same separation by color or sfr gives a systematic factor @xmath40 lower mass than separation by morphology , and @xmath19 is systematically higher than @xmath41 by about the same factor for all separation methods ( see also * ? ? ?
the systematic difference in @xmath41 and @xmath19 can be understood as a consequence of their definitions ( essentially one could define arbitrary schechter functions for early and late - type systems , and as long as the early - type function has a shallower slope and larger @xmath42 , this small systematic offset in the two measurements would be guaranteed ) .
the systematic offset between color / sfr and morphological selection is best demonstrated in the detailed comparison of local color and morphologically selected mfs in e.g. @xcite .
these authors find that although the different selection methods preserve the same qualitative behavior , and result in early and late - type samples which are identical in @xmath43 of the included galaxies , the color criterion does result in a slightly larger number of early - type systems ( probably owing to the large scatter in blue galaxy colors , with a non - negligible highly dust - reddened population ) , which will push the characteristic separation mass slightly lower .
this may also explain why there appears to be a larger scatter between samples at a given redshift in the morphologically - defined @xmath41 and @xmath19 , as such an effect will be sensitive to a given sample s resolution and imaging depth .
there may also be an interesting timescale effect , as discussed in @xcite , if galaxies redden onto the red sequence somewhat more rapidly than they morphologically relax following mergers ( perhaps suggesting different mechanisms for morphological and color transformation ) .
however , that @xmath44 ( depending only on @xmath45 ) is similar regardless of selection method suggests that these differences may be an artifact of the selection / identification of _ blue _ galaxies . in any case , further detailed study of these intermediate objects and comparison between different samples is needed to understand these differences .
these systematic distinctions caution against mixing definitions in determining the redshift evolution of these masses .
however , for our purposes , the systematic normalization scatter of a factor @xmath40 is not large this is comparable to the inherent ambiguity in defining a `` characteristic mass '' of any population ( e.g. schechter @xmath42 ) . as long as we are careful about the relatively small normalization offsets between selection criteria
, we can safely compare the `` transition '' mass and its evolution with redshift to the masses of other populations .
it has been suggested ( e.g. * ? ? ?
* ) that the `` transition '' mass may represent the mass at which the early - type mf is `` building up '' at each redshift , in the sense that `` new '' spheroids are being added to the rs mf at this mass . if cosmic downsizing applies to galaxy assembly to any extent
i.e. this `` building up '' extends at higher masses at higher redshift ( or `` peaks '' at lower redshift for lower masses ) , then this implies a shift of @xmath21 to lower mass at lower redshift @xcite .
to the extent that the shape of the late - type mf does not dramatically change over this redshift interval ( observed in the samples above and e.g. * ? ?
* ; * ? ? ?
* ) this will , by definition , manifest in a similar evolution of @xmath41 and @xmath19 . in this scenario ,
then , @xmath21 is directly tied to `` downsizing '' in the early - type mf .
figure [ fig : n.m.t ] considers the evolution with redshift of the number density of early - type galaxies of a given @xmath42 , from @xmath46 . at low @xmath42 ,
the number density declines steeply with @xmath13 ( roughly @xmath47 with @xmath48 ) .
although the statistics are poor and variance large at high @xmath42 , there is a significant trend ( @xmath49 ) for a shallower decline in number density at higher masses ( @xmath50 , i.e. little evolution in number density , at @xmath51 ) . above @xmath5 ,
the various samples plotted begin to disagree , and the @xmath47 functional form may not be a good approximation , so we reconsider this , fitting only the data at @xmath28 ( @xmath52 ) , and find the same trend at @xmath53 ( @xmath54 ) .
figure [ fig : ell.buildup.mfs ] considers the differential growth of the early - type mf in more detail .
we show the time - averaged buildup of early - type mfs in several redshift intervals from @xmath55 . where a given sample measures the elliptical mass function at two redshifts @xmath13 and @xmath56 , we differentiate the observed elliptical mass function at every mass with respect to the two redshifts to obtain the time - averaged rate of creation of ellipticals of that mass , over that redshift interval . in total , we show four redshift intervals : @xmath57 , @xmath58 , @xmath59 , and @xmath60 .
of course , we only compare mass functions measured with the same technique and sample in the same study , since systematic offsets in methodology could severely bias such an estimate .
although the scatter is large ( especially at low masses ) , the observations all trace a similar elliptical formation rate as a function of mass , with a similar break traced in the different samples at each redshift .
the shape of this function is not the same as that of the early type mass functions i.e. we are not simply recovering the fact that the mass function builds up uniformly over these intervals . at low redshift ( @xmath2 ) ,
comparison with the early - type mfs in figure [ fig : oplot.mfs ] shows that sub-@xmath42 ( @xmath61 ) ellipticals are `` building up '' in number density by @xmath62 per gyr , whereas the most massive systems build up by only @xmath63 per gyr . in other words
, the most massive systems are not building up ( either via star formation _ or _ assembly of stellar populations ) at a significant rate at low redshift . at high @xmath13
, there is a marginal shift of this function to higher masses .
by @xmath5 , comparison with the corresponding early - type mfs implies that systems with @xmath64 are building up by @xmath65 per gyr .
the best - fit schechter functions plotted in each redshift interval reflect this , with the schechter @xmath42 shifting from @xmath66 ( @xmath2 ) to @xmath67 ( @xmath68 ) .
the low-@xmath13 growth estimate of @xmath63 per gyr in the most massive systems is in excellent agreement with that from @xcite and @xcite determined from local red galaxy ( spheroid - spheroid or `` dry '' ) merger rates . in detail , observations suggest that the typical massive red galaxy undergoes @xmath69 major dissipationless mergers since @xmath70 @xcite . if this is representative , it is trivial to predict the corresponding rate of buildup of the elliptical population , assuming every red galaxy undergoes this number of major ( @xmath71 mass ratio ) mergers in this time .
this is shown in figure [ fig : ell.buildup.mfs ] , calculated from the local mf of @xcite .
although these mergers appear to be important for building up the most massive galaxies at low ( @xmath2 ) redshift , their contribution cuts off completely below @xmath72 ( and will cut off at higher masses at higher @xmath13 , generally about @xmath40 times the break in the `` buildup '' mass function ) .
such mergers move galaxies from the low - mass end of the elliptical mass function to the high mass end , and since the low - mass slope of the elliptical mass function is not steep , this can only _ decrease _ the number density of low - mass objects . by definition ,
then , dry mergers can not account for the ( substantial ) buildup of _ total _ mass on the rs nor the buildup at low and intermediate masses .
if these lower - mass ellipticals are formed by gas - rich mergers , then there must be at least @xmath40 times as many gas - rich mergers moving new galaxies onto the elliptical mass function as there are dry mergers ( in fact , the actual observed ratio is @xmath73 , @xcite ) .
figure [ fig : compare.mcrit.mstar ] compares the `` transition '' mass determined above with the characteristic masses ( schechter @xmath42 ) of the red , blue , and all galaxy populations . for clarity , we show just @xmath21 , as it is the most well - determined of the masses in figure [ fig : compare.methods ] , as well as being most robust with respect to sample definitions / selection , and further it has the most direct physical interpretation ( as it is not degenerate with blue cloud evolution ) . our conclusions here and subsequently , however , are unchanged regardless of the mass definition from figure [ fig : compare.methods ] .
the characteristic masses of red , blue , and all galaxies are nearly constant with redshift , with at most a marginal ( @xmath74dex ) increase from @xmath75 ; inconsistent with their following the strong trend seen in @xmath21 at @xmath76 ( independent of normalization ) . likewise
, comparing the shape of the rate of elliptical buildup in figure [ fig : ell.buildup.mfs ] with these galaxy mass functions at the same redshift shows that they do not trace the same mass distribution as a function of redshift .
we can therefore ( perhaps unsurprisingly , given the definitions employed ) rule out at high significance the hypothesis that `` transition '' mass objects are uniformly / randomly drawn from a `` parent '' population of normal galaxies of either early or late ( or both ) types .
it may appear that the strong trend in @xmath21 ( i.e. the mass above which the rs mf is @xmath77 assembled at @xmath13 ) is incompatible with the weak trend in @xmath42 of red galaxies ( as e.g. no change in @xmath42 would imply uniform buildup of rs populations at all masses ) .
this , however , is an artifact of the schechter function fit .
for example , given a local early - type schechter function mf with @xmath78 , @xmath79 , and @xmath80 and a similar mf at @xmath13 with @xmath81 , the same @xmath79 ( commonly assumed in fitting ) , and @xmath82 , one obtains @xmath83}},\ ] ] which is quite sensitive to @xmath84 and , for the observed values of @xmath85 , @xmath86 ( e.g. , * ? ? ?
* ) predicts a @xmath87dex evolution in @xmath21 similar to what we find . in a more physical sense ,
the local rs mf is the sum of the rs mfs built up over various intervals , each of which resembles a schechter function ( see figure [ fig : ell.buildup.mfs ] ) .
consider the sum of two schechter functions with identical @xmath79 and @xmath88 , but one having lower @xmath42 by @xmath89dex .
fitting this to a schechter function over the range @xmath90dex ( @xmath91dex ) yields a best - fit with @xmath42 only lowered from the higher value by @xmath92dex ( @xmath93dex ) .
this owes to the steep fall in @xmath17 at @xmath94 , which forces the combined fit to retain the high @xmath42 value . in further detail , if we imagine the rate of buildup @xmath95 ( figure [ fig : ell.buildup.mfs ] ) is a schechter function with constant @xmath79 and normalization , but an evolving @xmath96 ( adopting the best - fit trend @xmath44 shown in figure [ fig : compare.mcrit.mstar ] ) , then fit the integrated @xmath97 to a schechter function ( fitting over the range @xmath90dex ) , we obtain only @xmath98dex evolution in the early - type @xmath42 from @xmath8 to @xmath70 , despite the more than order - of - magnitude evolution in @xmath44 .
thus , although the strong evolution in the `` transition '' mass with redshift rules out its being representative of the general elliptical population , it is _ not _ inconsistent with the weak evolution in the early - type @xmath42 , even if @xmath21 does represent the characteristic mass at which `` new '' galaxies are being added to the red sequence . in other words , weak evolution in @xmath42 of red galaxies does not rule out strong evolution in the characteristic masses being `` built up '' on or added to the rs .
we next consider observed merger mfs .
we compile the local ( @xmath99 ) pair - selected major ( within @xmath100mag ) merger luminosity functions from @xcite in @xmath101-band and @xcite in @xmath32-band , as well as the morphologically identified merger / interacting galaxy luminosity functions from ( * ? ? ?
* cfrs+ldss ; @xmath102 ) in @xmath32-band and ( * ? ? ?
* gems+goods ; @xmath103 ) in the near - uv ( @xmath104 ) , and mass functions from ( * ? ? ?
* ; * ? ? ? *
hdf - n and hdf - s ; @xmath105 ) and ( * ? ? ?
* goods+deep2 ; @xmath106 ) .
where the mfs are not directly measured , we rescale the luminosity functions to mass functions using the mass - to - light ratios of ongoing mergers ( as a function of @xmath42 ) from ( * ? ? ?
* see table 1 ) .
these are calculated from the population synthesis models of @xcite , given the distribution of star formation histories during mergers determined from several hundred numerical simulations that include star formation , supernova feedback and metal enrichment , and black hole accretion and feedback ( see hopkins et al .
2006d for details ; this is essentially a second - order improvement on the typical empirically adopted simplified tau+burst models for these @xmath37 ) .
these should be reasonably robust : they have also been checked directly in the bands of interest here against the measurements of @xmath37 ratios in local ulirgs @xcite , pair samples @xcite , and recent merger remnants @xcite , and give good agreement ( * ? ? ?
* figures 1 & 4 ) .
furthermore , ( * ? ? ?
* figures 8 & 9 ) demonstrate that they can be reliably used to convert merger luminosity functions to mass functions ( in exactly this manner ) for all samples above where both are measured . in any case
, our subsequent results are unchanged ( albeit their significance reduced given the limited data ) if we consider only the morphologically identified , directly measured merger mass functions of @xcite and @xcite .
figure [ fig : merger.mfs ] shows several ( the most well - constrained ) of these mfs as a function of redshift , with the best - fit schechter functions . since constraints are weak at the faint end ( and systematic uncertainties large ; discussed below ) , we consider fits with fixed faint - end slopes @xmath79 , as well as allowing @xmath79 to vary . in all cases ,
there is a qualitatively similar trend for the schechter function @xmath42 to increase ( by @xmath107dex from @xmath75 ) .
direct comparison of the mfs demonstrates that this does not necessarily mean the number of mergers at the high - mass end increases monotonically with redshift ( changes in @xmath88 can offset the increase in @xmath42 ) ; this does however imply that the _ relative _ merger history / mfs favors higher mass mergers at higher redshifts .
although it is not important for our comparison , there is also a marginal trend for @xmath88 of mergers to increase with redshift relative to @xmath88 of the entire galaxy population , but this trend is largely driven by the low @xmath88 from @xcite .
figure [ fig : compare.mergers ] plots these schechter @xmath42 values , as well as those determined from the other merger mass / luminosity functions we compile , as a function of redshift , compared to the characteristic `` transition '' mass ( @xmath21 ) .
the characteristic masses from pair and morphologically - selected samples , as well as direct mfs , optical , near - ir , and near - uv luminosity functions agree surprisingly well at overlapping redshifts , at least up to the @xmath108 estimates which are strongly affected by cosmic variance @xcite , which further suggests the typical merger @xmath37 ratios used are reasonable .
there is a significant ( @xmath109 ) trend for the characteristic masses of mergers to increase with redshift . as with the `` transition ''
mass population in figure [ fig : compare.mcrit.mstar ] , this trend rules out at @xmath109 the hypothesis that mergers are randomly / uniformly drawn from the all or red galaxy population . whether mergers uniformly trace the blue galaxy population is less clear ; the values plotted are inconsistent with this hypothesis at @xmath110 , but the trend alone ( i.e. allowing for a systematic normalization offset ) is inconsistent at only @xmath111 .
the values / trend of the merger mf @xmath42 as a function of redshift are , however , similar and statistically consistent with @xmath21 ( and even more similar to @xmath12 and @xmath19 , see figure [ fig : compare.models ] ) .
we can consider in greater detail if observed merger mass / luminosity functions are consistent with the observed buildup of early - type populations by examining the complete mass functions as a function of redshift .
figure [ fig : ell.buildup.vs.merger.mfs ] reproduces figure [ fig : ell.buildup.mfs ] , but overlays the observed merger mfs from figure [ fig : merger.mfs ] at the appropriate observed redshifts . since there is considerable ambiguity in converting an observed merger mf to a merger
_ rate _ , we renormalize ( vertically only ; i.e. divide out an appropriate timescale ) the observed mfs arbitrarily such that we can focus here just on the unambiguous mass distribution ( although we will consider the issue of absolute rates / normalization shortly ) .
the agreement is striking : independent of the systematics in understanding merger rates , the observed distribution of mergers as a function of mass and redshift traces and is consistent with the buildup / addition of galaxies to the red sequence at all masses and redshifts observed .
although systematically uncertain , we should also compare the implied merger rates ( i.e. vertical normalization in figure [ fig : ell.buildup.vs.merger.mfs ] ) .
in other words , even if the buildup of early - type populations traces the merger mass distribution , are there the appropriate total number / rate of mergers to account for the growth of the rs mf ( assuming mergers are the agent of this `` buildup '' ) ? to estimate this , we assume every elliptical formation / addition event in figure [ fig : ell.buildup.mfs ] indeed owes to a merger , which is observable as such for some amount of time @xmath112 ( until morphological disturbances such as tidal tails fade beyond typical surface brightness limits ) . this yields the expected merger mf , @xmath113 . given the galaxy mfs from which we calculated these rates in the first place , it is then trivial to estimate the expected merger fraction as a function of mass .
figure [ fig : merger.frac ] compares this estimate , adopting a characteristic @xmath114 , with observed merger fractions as a function of mass and redshift .
this timescale is roughly expected from numerical simulations of mergers ( e.g. , * ? ? ?
* ) , dynamical friction considerations @xcite , or more detailed estimates of observational selection effects as a function of merger stage ( see , e.g. , * ? ? ?
* for a detailed discussion of these issues ) .
we consider the merger fraction above two representative stellar mass limits , @xmath115 and @xmath116 , but caution that these are not , for the most part , mass - selected samples , so the mass limits shown in figure [ fig : merger.frac ] are only broadly applicable .
we calculate the merger fraction as a function of mass directly from the merger mass functions of @xcite shown in figure [ fig : merger.mfs ] , at redshifts where the samples are complete to the given mass limit ( this essentially excludes the @xmath9 @xcite merger mass function ) .
the observed merger fractions are consistent with this estimate at all redshifts .
the buildup of elliptical populations does suggest that merger fractions should increase as a function of redshift , but we note that the effect is quite weak .
if , for example , the characteristic merger timescale decreases with redshift in the same manner as halo dynamical times ( at fixed mass ) , @xmath117 ( decreasing the expected merger fraction we calculate by this amount ) , the expected increase in merger fraction with redshift becomes marginal ( factor @xmath118 by @xmath119 ) .
both cases , however , are consistent with the present observations ( at these redshifts , within the factor @xmath40 typical uncertainties ) .
there is also marginal evidence for steeper evolution in merger fractions with redshift in the higher - mass cut we consider , evidence for which is also seen in e.g. @xcite , but we caution both that this trend appears only when different samples are combined , and that it will be `` washed out '' by the increasing importance of dry mergers at high masses and low redshifts ( see also figure [ fig : ell.buildup.vs.qso ] ) .
future studies which can separate gas - rich and gas - poor merger populations and track the merger fraction as a function of redshift and mass can provide a substantially stronger test of these trends .
caution regarding systematic uncertainties in merger populations is still warranted .
@xcite find that pair and morphological selection criteria yield similar results , but @xcite see significant disagreement , which may be the result of systematic identification of `` false '' ( i.e. non - merging ) pairs at low redshift @xcite .
fortunately , the characteristic merger mass ( or luminosity ) does not appear to change dramatically with selection method even though the time spent in a given phase ( and thus @xmath88 or merger fraction ) may .
for example , the data of @xcite and @xcite do yield a similar characteristic merging galaxy luminosity @xmath120 , despite finding different merger fractions .
one might also wonder whether the natural tendency of a flux - limited sample to select brighter systems at higher redshift might lead one to infer an increasing mass scale regardless of the underlying mass distribution .
however , figure [ fig : merger.mfs ] demonstrates that the completeness limits for most of the samples we consider are generally well below the `` transition '' mass , and similarly below the break in the corresponding mass function .
further , although @xcite find that the observed number of faint mergers depends on selection effects ( see figure [ fig : ell.buildup.vs.merger.mfs ] ; which shows the increase in number of low - mass mergers when increasing imaging depth and decreasing field size by an order of magnitude , from gems to goods ) , this does not significantly effect the merger @xmath42 or change our comparison in figure [ fig : ell.buildup.vs.merger.mfs ] .
still , these effects must be accounted for in any comparison of fitted mfs .
it is furthermore true that the exact appropriate value of the duration of observable merger activity ( @xmath112 ) is not well - determined , and will in detail depend on the sample , mass limit , and redshift , but for our purposes these effects ( amounting to a systematic factor @xmath40 uncertainty ) are generally comparable to or smaller than the scatter in the observations in figure [ fig : merger.frac ] .
preliminary estimates of the observable merger timescale based on comparison with automated nonparametric classification schemes @xcite suggest , perhaps surprisingly , relatively weak trends with redshift ( at least at @xmath11 ; see the discussion in hopkins et al .
2006d ) , but lacking a complete cosmological framework from which to predict observable merger properties , figure [ fig : merger.frac ] should be taken with the strong caveat that the relative normalizations of galaxy buildup and observed mergers depends systematically on @xmath112 .
detailed studies of `` e+a '' ( or `` k+a '' ) galaxies @xcite , with characteristic post - starburst stellar populations indicating a substantial but rapidly quenched star formation epoch in the last @xmath121gyr @xcite , have found ubiquitous evidence of morphological disturbances and tidal tails @xcite , which together with their environmental ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and structural / kinematic properties @xcite imply their formation in mergers and evolution into typical early - type galaxies .
we therefore consider whether e+a galaxies , presumably recently formed in mergers , trace any `` downsizing '' trend .
figure [ fig : compare.ea ] compares @xmath21 and the masses of observed e+a galaxies as a function of redshift . at low redshift , sizable samples exist , and we show the characteristic schechter function @xmath42 of e+a populations . at higher redshift , samples are extremely limited , and we can only plot the masses of individual systems .
the points as plotted appear to `` downsize , '' as noted in @xcite , but this trend could well be completely driven by survey flux limits .
lacking volume - limited samples or complete e+a mfs at high redshift , we can only presently say that the e+a data are not _ inconsistent _ with the downsizing in the `` transition '' mass or any of the other hypotheses considered herein .
if the formation / movement of galaxies on the rs is associated with a quasar `` trigger '' ( for example , through quasar feedback being an agent of reddening , or both being associated with a merger ) , then the observed quasar luminosity function ( qlf ) should reflect the rate of elliptical formation / buildup shown in figure [ fig : ell.buildup.mfs ] indeed , in such a model , each quasar `` broadcasts '' a galaxy moving to / forming on the red sequence .
( if the quasar `` lifetime '' were of order the hubble time , of course , then the qlf would reflect the integrated / established early - type population , but observations constrain it to be much less at all redshifts , e.g. martini 2004 ; hopkins , narayan , & hernquist 2006 ; such that the qlf tracks the rate of `` triggering . '' ) it is straightforward to compare these .
a spheroid of mass @xmath42 hosts a black hole of mass @xmath122 ( @xmath123 ; * ? ? ?
* ; * ? ? ?
* ) , confirmed by direct observations at all redshifts of interest ( @xmath11 ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
a `` quasar '' event is essentially defined by `` ignition '' of the black hole for a brief time ( @xmath124yr from various observations ; see * ? ? ? * and references therein ) near the eddington limit @xmath125 ( where @xmath126 ; e.g. , * ? ? ?
* ; * ? ? ?
thus , the formation or movement to the rs of a spheroid of mass @xmath42 would be associated , in this scenario , with a short - lived quasar of luminosity @xmath127 this simple effective conversion for bright quasars is supported by both numerical simulations of quasars and galaxy mergers @xcite and direct comparison of quasar and host galaxy luminosities @xcite .
there is , of course , some uncertainty and observed scatter in the host galaxy - bh mass correlation and bright quasar eddington ratios , but it is constrained to a factor @xmath40 , comparable to the uncertainty in the observed @xmath12 .
if this is the dominant mode of quasar triggering , then although the exact normalization of the qlf ( number of observed quasars ) will depend on the `` duty cycle '' @xmath84 and quasar lifetime @xmath128 , the break @xmath42 in the host / source mass function ( break in the rate of `` formation '' of early - type galaxies calculated in figure [ fig : ell.buildup.mfs ] ) will translate directly to a break @xmath120 ( from equation [ eqn : ledd ] ) in the qlf .
if @xmath84 and/or @xmath128 are complicated functions of mass , luminosity , or redshift , they might change the slopes of the resulting qlf , but will not move the break location @xmath120 .
thus , @xmath120 directly tracks the characteristic mass of the host population .
put another way : essentially all observed @xmath129 quasars have @xmath130 , i.e. define a characteristic active black hole mass @xmath131 , and since @xmath132 at all redshifts of interest , the characteristic host mass @xmath42 of quasars is well - defined at each @xmath13 .
figure [ fig : qlf.break ] compares the qlf characteristic luminosity expected ( equation [ eqn : ledd ] ) from objects of the `` transition '' mass ( @xmath21 ) and the observed characteristic qlf `` break '' luminosity @xmath120 as a function of redshift .
the break @xmath120 is determined in the standard fashion , fitting the observed qlf at each redshift to a double power - law .
we show @xmath120 measured from optical , soft x - ray , and hard x - ray studies , each converted ( to enable direct comparison ) to a bolometric luminosity @xmath133 using a standard observationally - derived bolometric correction ( template quasar sed ) and reddening correction ( * ? ? ?
* and references therein ; note that adopting the less recent bolometric corrections from elvis et al .
1994 or marconi et al .
2004 yields nearly identical results ) .
we also show the break determined by ( * ? ? ?
* see their table 2 ) , who compile a large number of qlf measurements through the mid and near - ir , optical , near - uv , soft and hard x - ray , and soft gamma ray and use these to directly determine the bolometric qlf . in any case
, the observed @xmath120 is robust ; in fact , the ( typical factor @xmath40 ) discrepancies in @xmath120 owe mostly to the data binning and fitting function , and a direct comparison of the data in @xcite shows they trace a similar turnover / break .
we also note that the existence of a break is unambiguous ( detected at @xmath134 in most of the samples in figure [ fig : qlf.break ] ) , regardless of whether it is sharp ( as expected for a double power - law fit ) or exhibits some higher - order curvature ( as for a schechter function ; e.g. , * ? ? ?
* ; * ? ? ?
it is unaffected by questions of completeness , as the x - ray surveys are typically complete to @xmath40 orders of magnitude in luminosity below @xmath120 ( e.g. , * ? ? ?
the break luminosity also increases with redshift ( at least to @xmath4 ) , as has long been recognized in quasar surveys ( recently , e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* significant in each case at @xmath135 ) , regardless of higher - order subtleties implied by `` luminosity - dependent density evolution '' ( e.g. , * ? ? ?
* and references therein ) models and other changes in the detailed qlf shape as a function of redshift ( e.g. , * ? ? ?
* ; * ? ? ?
that the expected @xmath120 from @xmath21 agrees with the observations is not surprising .
@xcite combined observed merger mass functions with a large suite of corresponding hydrodynamical merger simulations @xcite including black hole accretion and feedback , to calculate what the resulting merger - driven qlf should be in each band at each corresponding redshift .
their predicted merger - driven qlf agrees well ( @xmath136 ) with that observed at every redshift at which this comparison is possible , and the inverse ( predicted merger mf from the qlf ) also agrees well with the observations . since we have shown that the buildup of the elliptical mf and @xmath21 trace a similar mass distribution to mergers , it is expected that the resulting @xmath120 will agree with the qlf .
however , our comparison here , unlike @xcite , is model - independent , based only on the well - determined black hole - host mass relation and eddington limit . in figure [
fig : compare.quasars ] we invert this comparison , and estimate the characteristic `` transition '' mass expected based on the observed qlf .
we first show points as in figure [ fig : qlf.break ] , estimating a characteristic host mass from the qlf @xmath120 ( inverting equation [ eqn : ledd ] ) .
however , a proper calculation is not so trivial , as in detail @xmath21 will be determined by the integrated buildup of the early - type mf ( and late - type mf for @xmath12 , @xmath19 ) , and therefore requires that we adopt some model for quasar light curves and triggering .
hopkins et al .
( 2006a , b ) use a large set of several hundred hydrodynamical simulations @xcite of galaxy mergers , varying the relevant physics , galaxy properties , orbits , and system masses , to quantify the quasar lifetime ( and related statistics ) as a function of the quasar luminosity .
they define the quantity @xmath137 , i.e. the time a quasar of a given bh mass ( equivalently , peak quasar luminosity @xmath138 ) will be observed at a given luminosity @xmath133 .
they further demonstrate that this quantity is robust across the wide range of varied physics and merger properties ; for example , to the extent that the final bh mass is the same , any `` major '' merger of sufficient mass ratio ( less than @xmath139 ) will produce an identical effect .
given the tight black hole - host mass relation ( @xmath140 ) , it is trivial to write this as @xmath141 . since at all @xmath133 , @xmath142 ( the hubble time )
, the observed qlf @xmath143 is given by @xmath144 where @xmath145 ( hereafter @xmath146 ) is the rate of quasar `` triggering '' as a function of host spheroid mass at a given redshift . if the `` trigger '' is associated with formation of the spheroid or movement of a galaxy to the rs , then @xmath147 _ is _ the rate of buildup of the rs as a function of mass , directly comparable to that in figure [ fig : ell.buildup.mfs ] .
knowing @xmath143 directly from observations and adopting the @xcite @xmath141 ( which is at least consistent with all quasar observational constraints ) , the inversion of equation ( [ eqn : qlf ] ) yields @xmath147 .
@xcite perform this inversion , using their large compilation of observed @xmath143 , and quote the best - fit @xmath147 ( table 5 ) .
we adopt their best - fit model for each redshift interval ( individually ; although a global fit yields similar results ) .
if , again , this represents the buildup of the rs , then we can integrate from @xmath148 to obtain the rs mf at all redshifts and calculate @xmath44 .
the late - type mf is reasonably well - measured over the range of interest , so comparing it with this integration also yields an expected @xmath12 and @xmath19 .
we compare these estimates with the observed @xmath21 in figure [ fig : compare.quasars ] , and find they agree at all observed redshifts .
having obtained the rate of `` buildup '' of early type mfs expected if each quasar `` trigger '' is associated with the formation / movement of a rs galaxy , we can directly compare with the rate of buildup implied by observed early - type mfs .
figure [ fig : ell.buildup.vs.qso ] plots the time - averaged buildup determined from the qlf , from the same @xmath147 as figure [ fig : compare.quasars ] , compared with the observed buildup from figure [ fig : ell.buildup.mfs ] .
we consider both the mean time - averaged buildup ( averaged over each appropriate redshift interval ) assuming each quasar `` trigger '' is instantaneously associated with the movement of a galaxy to the rs , and that expected if there is a uniform @xmath100gyr delay after each quasar before the galaxy becomes red ( allowing time for e.g. gas exhaustion and reddening ) . in either case , this estimate agrees with the observed `` buildup '' of elliptical populations , at all masses and redshifts ( @xmath149 at all @xmath150 ) .
the latter ( @xmath151gyr delay ) case gives marginally better agreement , but the difference between the two is comparable to the uncertainties in either determination of @xmath147 ( see figure [ fig : compare.quasars ] ) . at the highest masses at low redshifts ( @xmath2
; although also to a lesser extent at @xmath152 ) , this estimate falls short of observed rates of `` buildup . ''
however , this is precisely where we have estimated that observed rates of `` dry '' mergers can account for early - type growth .
since gas - free mergers are not expected to trigger quasar activity , it is not surprising that this would not be implicit in quasar luminosity functions .
allowing for the contribution of dry mergers shown at @xmath2 improves the agreement considerably ( @xmath153 ) .
there is still some tension matching the observations near @xmath154 , but it is important to note that at these redshifts , the cosmic variance associated with small volume , narrow - field galaxy surveys and , perhaps more importantly , with even wide - field quasar surveys ( given the very low local space density of quasars ) , is largest .
having estimated the rate of quasar `` triggers , '' @xmath147 , in figure [ fig : ell.buildup.vs.qso ] , then if each such `` trigger '' is in fact a galaxy merger , we can convert this to an expected merger fraction in exactly the same manner as we converted the rate of early - type `` buildup '' in figure [ fig : ell.buildup.mfs ] ( i.e. simply assuming an observable merger timescale @xmath112 ) .
this is shown in figure [ fig : merger.frac ] , along with the observed additional contribution from dry mergers .
given the agreement with the rate of elliptical `` buildup '' in figure [ fig : ell.buildup.vs.qso ] , it is not surprising to find this agrees with observed merger fractions .
as a caution , we should note that these calculations can give a misleading result if the full luminosity dependence of the quasar `` lifetime '' from simulations @xcite and observations ( e.g. , * ? ? ? * ; * ? ? ?
* ) is not properly taken into account .
such a case is not , of course , well - motivated physically , although it may represent alternative quasar feedback models ( or a complete lack of such feedback ) , but it is nevertheless sometimes adopted for simplicity .
why should such a simplified model give a qualitatively different result ?
in the @xmath137 model we consider , @xmath128 is larger at low luminosities , because low - level agn activity can persist for a long time after the violent , sudden high - accretion rate episode in a merger . ignoring this luminosity dependence and assuming , e.g. that all quasars turn on and off ( as `` light - bulbs '' ) for a short time implies that all observed quasars , even those at very low luminosity , are seen at ( or very near ) their `` trigger , '' i.e. are in _ ongoing _ mergers .
this gives a misleading estimate of the number of mergers needed to account for the qlf , and as a result yields an incorrect estimate of host luminosity functions and black hole mass functions ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , as well as , consequently , erroneous estimates of the associated `` transition '' mass . having considered the @xcite models of merger - triggered quasar lightcurves , we briefly note additional future tests of these models and the generic association between the blue - red transition or elliptical formation and quasar activity . in figure
[ fig : quasar.tests ] we compare the observed qlf with the expected `` conditional '' qlf , i.e. the contribution to the qlf from hosts / merger remnants with different masses relative to the observed `` transition '' mass . in other words , the contribution @xmath155 from equation ( [ eqn : qlf ] ) .
the qlf near @xmath120 corresponds to objects with @xmath156 . at the faintest and brightest luminosities
, there are contributions from smaller and larger hosts , respectively ( and a significant fraction of objects at the lowest luminosities will not necessarily be merger - triggered ; @xcite ) , but it is clear from the figure that a direct measurement of the host masses of characteristic quasars at @xmath13 should find their hosts dominated by objects with @xmath157 or @xmath21 , many of which should appear as relatively young ellipticals , if this picture is correct . in this scenario , one might also expect that objects around the `` transition '' mass preferentially host agn .
since equation ( [ eqn : qlf ] ) implicitly defines the probability to see a galaxy with mass @xmath42 ( bh mass @xmath158 ) at luminosity @xmath133 , we can estimate the fraction of such galaxies with a given eddington ratio .
figure [ fig : quasar.tests ] plots the fraction of galaxies hosting an agn with an accretion rate ( relative to eddington ) @xmath159 as a function of galaxy stellar mass at several redshifts , compared to the observed `` transition '' mass at that @xmath13 .
( note the actual agn luminosity will be @xmath160 . )
this particular prediction is sensitive to the different fits to the `` triggering '' rate @xmath147 provided in @xcite , especially at low mass ( @xmath161 ) , but the trend that the peak / turnover in this distribution tracks the `` transition '' mass is robust .
in most semi - analytic models , gas infalling in dark matter halos is shock - heated to the virial temperature , and , in low mass halos , subsequently cools on a short timescale , allowing rapid accretion onto the central halo galaxy and defining a `` rapid cooling '' or `` cold accretion '' regime .
however , in massive halos , the cooling time is longer and gas forms a quasi - static `` hot '' halo , defining a `` static hot halo '' or `` hot accretion '' regime ( e.g. , * ? ? ?
* ; * ? ? ?
more recently , it has been suggested that the transition between these regimes is sharp , near a halo mass @xmath162 ( although this number is uncertain by a factor of several ) at low redshift @xcite , and that suppression of future cooling and accretion is very efficient , essentially `` cutting off '' all gas supplies above this mass ( e.g. , * ? ? ? * ) and `` quenching '' star formation .
the shock - heating of infalling gas need not be the _ specific _ physical agent of this `` quenching '' : for example `` radio mode '' or low - luminosity , continuous agn feedback @xcite or cyclic , short - lived quasar activity @xcite may be invoked to maintain the gas in the `` hot '' phase .
there is therefore a potentially important distinction between semi - analytic models ( sams ) which assume that the feedback mechanism is `` at ready , '' such that upon crossing the critical mass @xmath163 , star formation and gas accretion onto the central galaxy is _ instantaneously _ terminated , and those that require some additional mechanism or process ( such as the formation of a relatively massive bulge and black hole ) to drive the blue - red transition and transformation / movement of galaxies to the rs .
this essentially relates to the important distinction , discussed in [ sec : intro ] , between the mechanism by which galaxies _ become _ red / elliptical and that by which they _ maintain _ their colors / low star formation rates .
the key value of invoking this `` hot accretion '' regime in sams has been the ability to suppress star formation on timescales of order the hubble time .
however , although this could , in principle , be _ necessary _ to yield red galaxies at @xmath8 , it does not automatically follow that it is _ sufficient_. in other words , there may be other processes ( e.g. mergers and/or quasars ) which drive the blue - red transition and movement to the rs , and the `` hot accretion '' mode simply maintains these galaxies at their low star formation rates .
one possible interpretation of the observed `` transition mass , '' perhaps the most naive , is that the transition mass simply represents the stellar mass hosted in @xmath163 halos at each redshift .
if we adopt the expected halo quenching mass @xmath163 from @xcite , and either assume the galaxies hosted have the same stellar mass as those in @xmath8 halos of the same mass ( measured in * ? ?
* ) , or that they are already fully assembled ( i.e. have stellar masses at @xmath13 appropriate for what their halo mass will be at @xmath8 ) , we can compare with our observed `` transition mass . ''
we find that while the two are similar at low redshifts , they diverge at higher-@xmath13 .
this is , of course , where the observations are most uncertain , so it may simple reflect a systematic error in our estimation of the `` transition mass . ''
but it probably also reflects the possibility that , in these models , the `` transition mass '' has a more complex physical origin than simply tracing @xmath163 . as noted in
[ sec : compare.mstar ] , allowing for more complex and realistic distributions of galaxies in transition to the rs can affect quantities such as the `` break '' mass @xmath42 and `` transition '' mass in a non - trivial manner .
but there is also the possibility that the `` transition '' to the rs requires additional processes beyond the initial cutoff of new gas supplies in `` hot mode '' accretion , such as gas exhaustion , mergers , and/or quasar activity to operate , which are what we see traced by the observed transition mass .
it is also worth considering whether or not the flow of galaxy host halos across @xmath163 is consistent with the number and mass densities of `` transition '' objects and the buildup in early - type populations which we have estimted from the observations .
the rate at which halos cross a given mass threshold @xmath164 is straightforward to calculate in linear theory from the press - schechter formalism , @xmath165 and adopting either a simple mean stellar - to - dark matter mass ratio ( from the calibration of * ? ? ?
* ) , or integrating ( at this halo mass ) over the population of inferred hosted galaxies from the conditional galaxy mass functions ( cmfs ; i.e. probability that halos of mass @xmath164 host galaxies of stellar mass @xmath42 ) yields an estimate of the rate at which stellar mass crosses this threshold .
the local cmf is determined ( albeit indirectly ) entirely from observations of galaxy mass / luminosity functions and clustering ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and has been subsequently measured directly in 2dfgrs group catalogues by @xcite , and is well - constrained with typical uncertainties smaller than or comparable to those in our estimate of the rate of early - type `` buildup '' ( at least for @xmath166 of interest here ; the mf at lower masses depends on the mass threshold for inclusion of satellite systems ) .
@xcite extend the clf / cmf to high redshifts ( @xmath167 ) using a large number of luminosity function and clustering estimates from wide - area surveys .
note that @xcite and @xcite actually measure the conditional luminosity function ( clf ) ; we convert to a cmf using the appropriate @xmath37 ratios as a function of mass from @xcite , and assume these @xmath37 values evolve with redshift following the best - fit stellar population models as a function of mass from @xcite and @xcite .
checking directly ( following the methodology of * ? ? ?
* ) shows that this agrees with the @xcite mass functions ( see also their figure 19 ) and the @xcite measured clustering as a function of stellar mass , and furthermore , these cmfs agree well with those directly determined in @xcite .
ultimately , there are a number of systematic ( factor @xmath168 ) uncertainties in this comparison , and our ( admittedly crude ) empirical calculation ignores the fact that , in `` quenching '' models , crossing the quenching threshold itself may change the stellar - to - dark matter mass ratios and stellar @xmath37 values .
however , within these rather large uncertainties , our purely empirically estimated rate at which galactic host halos cross @xmath163 is consistent with the possibility that this is a necessary prerequisite for `` transition '' to the rs .
we compare the populations we have considered in an independent manner by examining their clustering properties . if a population ( i.e. a given set of `` parent '' halos ) clusters with a given bias @xmath169 at some redshift @xmath13 , then the subsequent evolution in their bias is trivially calculated in linear theory @xmath170 , \label{eqn : bias}\ ] ] where @xmath171 is the growth factor @xcite , regardless of the processes ( accretion , mergers , etc . ) that affect the halos ( and galaxies ) themselves .
the bias of galaxies ( specifically red / elliptical galaxies ) as a function of stellar mass is well - determined at @xmath8 ( we adopt the recent determination from the sdss in * ? ? ? * with typical @xmath172 uncertainty ) , so given the bias of a population at @xmath13 and evolving it to @xmath8 with equation ( [ eqn : bias ] ) yields the characteristic @xmath8 stellar mass of this population ( i.e. the average stellar mass of which the population is the `` parent '' ) .
figure [ fig : clustering ] shows this @xmath42 , calculated from various clustering measurements @xmath169 of quasars , mergers ( ulirgs and smgs ) , and e+a galaxies as a function of redshift , and compares to the `` transition '' mass @xmath21 at each redshift .
we invert this as well ; knowing @xmath44 , evolve @xmath173 with equation ( [ eqn : bias ] ) to estimate @xmath169 .
note that all @xmath169 shown from measurements are converted from the directly observed clustering length @xmath174 , which for a power law correlation function yields @xmath175^{\gamma/2},\ ] ] with @xmath176 and @xmath177 ( see , e.g. , * ? ? ?
the absolute value of the bias as plotted is then weakly dependent on cosmology ( and this conversion , of course , is inexact ) , but the important point for our purposes is that the _ relative _ bias of all points plotted ( and @xmath178 with which we compare ) is insensitive to the cosmology .
these comparisons do assume that the stellar mass of individual systems does not change much from @xmath13 to @xmath8 , i.e. that , once formed , ellipticals are passively evolving .
however , we can easily eliminate this assumption , by considering the clustering directly observed for red galaxies of mass @xmath44 at that redshift , and figure [ fig : clustering ] shows this as well . in either case , the agreement with the clustering of quasars and ( albeit much less well - constrained ) merger / e+a populations is good .
this also agrees with determinations of e.g. the typical overdensities and small - scale clustering of quasars and ulirgs @xcite .
this method by which we compare clustering is only weakly dependent on cosmology , through the growth factor @xmath171 ( independent of e.g. @xmath179 ) .
there are some caveats , however .
technically , we are estimating the mass which has exactly the observed bias ; this is some weighted mean mass .
however , theoretical expectations from physically motivated quasar light curve models @xcite and direct observations of clustering as a function of luminosity @xcite suggest that quasar clustering depends only weakly on luminosity , reflecting a reasonably well - defined characteristic host mass .
these comparisons will also , of course , be affected if the clustering of mergers is different on large scales from that of non - merging halos of the same mass ( a so - called `` merger bias '' ) . however , a number of investigations have found no such dependence ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and even where more recent investigations have seen such an effect @xcite it has been restricted to small mass halos ( below the `` collapse mass '' , i.e. where @xmath180 ) at @xmath8 , and therefore the assumption of no merger bias has generally been adopted in quasar clustering studies @xcite .
we can repeat this comparison using the formalism of @xcite from linear collapse theory , which yields a characteristic halo mass from a given observed @xmath169 .
we use the observed stellar mass - halo mass relations calibrated for elliptical galaxies from weak lensing measurements in @xcite to convert these halo masses to a stellar mass @xmath41 .
we convert between halo mass and bias with the method of @xcite modified following @xcite in our adopted cosmology ( in detail assuming @xmath181 , @xmath182 ) with the power spectrum computed following @xcite .
the results are similar , but are much more sensitive to the adopted cosmology and systematics in the stellar mass - halo mass relation in this approach . in considering the clustering of @xmath21 objects , we have considered the directly measured bias of objects with mass @xmath44 at redshift @xmath13 , as well as the `` passively evolved '' clustering from the @xmath8 bias as a function of mass .
we can gain further insight into the evolution of these populations by comparing the two . knowing the observed bias of @xmath44 objects at @xmath13
, we can evolve this to @xmath8 given equation ( [ eqn : bias ] ) , and then use @xmath183 to obtain the typical stellar mass hosted by these systems _ at @xmath8 .
_ comparing that to their stellar mass at @xmath13 , namely @xmath44 , shows by how much the typical stellar mass of the population has grown .
we could also estimate this in a more indirect fashion , using linear theory to estimate a host halo mass @xmath184 given @xmath169 , then knowing the @xmath8 mass of a halo with mass @xmath184 at @xmath13 , use the local galaxy stellar - halo mass calibrations from @xcite to obtain @xmath185 .
again , this approach is considerably more sensitive to the assumed cosmology , but in our adopted case yields similar results . in figure
[ fig : clustering.to.m.z0 ] , we use this to compare @xmath44 , the stellar mass of `` transition '' mass objects at @xmath13 , with @xmath185 , i.e. the mean @xmath8 stellar mass which is typically hosted by the evolved `` parent halos . '' unless `` transition '' mass objects comprise some unusual outlier in their halo properties , this should represent the typical stellar mass these objects will grow to by @xmath8 .
we compare with the expectation , following e.g. @xcite , that these stellar masses grow at a rate corresponding to one major ( mass ratio @xmath186 ) merger since @xmath70 .
we also consider the case if the stellar mass in these objects grows in fixed proportion with their host dark matter halos .
the systematic uncertainties ( and measurement errors in @xmath169 ) are sufficiently large that we should regard these comparisons with caution , and _ not _ consider this as evidence for a particular `` amount '' of dry merging .
however , the estimated @xmath185 demonstrates that the @xmath44 measurements are completely _ consistent _ with subsequent growth by dry mergers at observationally inferred rates .
growth in proportion to the host halo mass , by contrast , is extremely difficult to reconcile with observed properties of the galaxies .
this is not surprising , as the existence of any significant @xmath187 galaxy population at @xmath70 without a corresponding @xmath188 galaxy population at @xmath8 implies that , at least for some objects assembled most rapidly , subsequent galaxy _ assembly _ must lag behind halo growth ( or subsequent growth in these halos must be anomalously slow ) .
this does mean , however , that it is _ not _ possible to reconcile the observations with a model in which galaxy assembly uniformly tracks halo assembly , even allowing for the final galaxy stellar - halo mass ratio to be a function of halo mass ( i.e. setting in all progenitors the effective @xmath37 of the @xmath8 halo , which then simply assembles ) .
we compile a large number of observations of red / elliptical galaxy mass functions , and use these to determine the rate of `` buildup '' of the red sequence ( rs ) as a function of mass and redshift .
comparing these with observations of other populations allows us to test a number of different models for the possible associations between these populations and the `` transition '' of galaxies from blue , star - forming disks to red , `` dead '' ellipticals .
independent of the nature of `` downsizing '' in the buildup of rs mass functions ( discussed below ) , the rate of rs `` buildup '' is sufficiently well - determined to place meaningful constraints on a number of models .
dissipationless ( gas poor , red , or `` dry '' ) mergers can account for the buildup of the rs at only the largest masses @xmath1 at low redshift ( @xmath2 ) . at higher redshifts ( @xmath189 ) ,
the dry merger rate would have to be at least an order of magnitude larger than observationally estimated @xcite to account for observed rs buildup , even at the highest masses .
this is perhaps unsurprising , as these and other observations find the _ gas - rich _ merger rate / fraction is an order of magnitude or more larger at all but the lowest redshifts .
furthermore , the total mass density on the rs is observed to increase by a factor @xmath190 since @xmath5 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and dry mergers can not , by definition , move / form `` new '' galaxies and mass on the rs
. however , we find the _ total _ observed merger population ( gas rich+gas poor ) agrees very well with that expected if all rs galaxies are formed in mergers . both the detailed mass distribution and fraction / rates of galaxy mergers are consistent with the rate of rs buildup at all masses and redshifts observed .
this merger population is dominated by gas - rich mergers at all masses at high redshifts ( @xmath189 * ? ? ?
* ; * ? ? ?
* ) and at low masses at low redshifts , morphologically identifiable as bright ( i.e. star - forming or starbursting ) interacting systems ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . in detail , completely neglecting dry mergers ( or merger mass functions sensitive to them
) , this agreement is unchanged except for the highest masses at low-@xmath13 discussed above .
there is substantial systematic uncertainty in converting a merger fraction to a merger rate ; our comparisons assume a characteristic observable merger timescale @xmath191gyr .
however , this is a theoretically reasonable timescale ( see [ sec : compare.merger.mass ] ) , and given the scatter in the observations , our conclusions are not changed for systematic shifts within a factor @xmath118 , nor for allowing the merger timescale to scale with halo dynamical times ( @xmath117 ) .
furthermore , this has no effect our comparison of the mass distributions of these populations .
similarly , we find the rate at which host galaxies trigger quasars , determined as a function of the host stellar mass and redshift from the quasar luminosity function , agrees well with the observed rs buildup at all masses and redshifts observed .
there is some discrepancy at the lowest redshifts and highest masses , but this is again where the dry merger contribution can account for the observed buildup , and dry mergers ( by definition being gas - poor or gas - free ) are not expected to trigger quasar activity .
we consider this comparison first in a purely empirical fashion , using observed quasar eddington ratios and the black hole - host mass relation to estimate quasar host masses as a function of redshift , and then in greater detail adopting the models of quasar light curves and lifetimes as a function of luminosity and host properties from the simulations of merger - induced quasar activity in @xcite .
the latter introduces some model dependence ( although it is consistent with the eddington ratio and black hole - host mass relation estimates ) , but allows us to consider this comparison in greater detail and to make specific predictions for the characteristic host masses of quasars as a function of their position on the qlf and for the agn or `` active '' fraction of galaxies as a function of stellar mass . in either case , the agreement between the rates of quasar formation / triggering as a function of host stellar mass and the buildup of rs galaxies is similar .
we independently test these possible associations by comparing clustering measurements of the relevant populations as a function of redshift , and find similar results .
the clustering of quasars and systems `` in transition '' to the rs agree at all redshifts as if they trace the same mass distribution .
clustering properties of merger ( ulirg and smg ) and post - merger ( e+a ) populations are consistent , but considerably less well - constrained .
although the above comparisons do not technically depend on it , we determine the `` transition '' mass ( @xmath12 , @xmath19 ) , i.e. the mass which separates the blue , star - forming disk and red , non star - forming elliptical populations , as a function of redshift .
it has been suggested ( e.g. , * ? ? ?
* ) that this represents the characteristic mass at which galaxies are forming on or being added to the rs as a function of redshift , but quantified in this manner , it is not obviously so ( see , e.g. * ? ? ?
we therefore also determine ( @xmath21 , following * ? ? ?
* ) the minimum mass above which the rs mass function is @xmath3 assembled at a given redshift .
regardless of definition , and furthermore regardless of the criterion used to separate early and late - type populations ( whether e.g. a color , star formation rate , or morphology criterion ) , @xmath12/@xmath19/@xmath21 shift to systematically larger masses at higher redshift ( significant at @xmath192 ) , tracing a very similar trend as a function of redshift .
this trend , especially in @xmath21 ( which is independent of possible evolution in late - type mass functions ) suggests that `` downsizing '' applies not just to galaxy _ star formation _ , but also in some sense to galaxy _ assembly _ , as suggested by the studies of e.g. @xcite . in greater detail , considering the full rate of rs buildup as a function of stellar mass and redshift , low mass ( @xmath193 )
galaxies appear to be building up rapidly / continuously at low redshifts ( @xmath62 per gyr ) , but the most massive systems do not ( @xmath63 per gyr growth at @xmath194 ) .
the growth of the most massive systems instead appears to be rapid at significantly higher redshifts ( e.g. rising to @xmath65 per gyr by @xmath5 ) .
equivalently , the characteristic mass ( schechter function @xmath42 ) defined by this `` formation '' rate appears ( albeit at only @xmath195 ) to increase with redshift in a similar fashion to the `` transition '' mass .
we compare the `` transition '' mass with the characteristic masses of mergers and quasars and again find they trace similar masses as a function of redshift , with `` downsizing '' evident in all three populations ( @xmath109 for mergers , @xmath192 for quasars ) , further supported by their observed clustering .
we compare with the characteristic ( schechter function ) @xmath42 of the entire , red , and blue galaxy populations , and rule out at high significance the possibility that `` transition '' mass objects are drawn uniformly from any of these populations .
even with the systematic uncertainties in this mass estimate , it is also clearly distinct as a function of redshift from the characteristic masses of e.g. cluster , radio galaxy , ero , drg , or lbg populations ( see e.g. figure 2 of * ? ? ?
these observations are all consistent with and suggest a scenario in which major mergers , quasars , and the transition from blue disk to red elliptical galaxies are associated .
they do not inform us regarding , for example , whether gas exhaustion or stellar or quasar feedback is the specific mechanism for the reddening which accompanies the merger - driven morphological transformation and quasar episode .
however , they support the hypothesis that mergers drive the transition from blue disks to red elliptical galaxies , terminating in decaying , feedback - driven bright quasar phases .
the transition mass and break in the qlf appear to reflect the characteristic mass of gas - rich objects merging at a given redshift , which may build up the _ new mass _ on the red sequence at progressively lower masses at lower @xmath13 as gas supplies are exhausted in more massive systems . that quasar host masses trace the `` transition '' mass and not , e.g. the blue galaxy population @xmath42 ( see also figure [ fig : compare.models ] ) rules out the possibility that quasar activity _ generically _ traces star formation , as variants of e.g. the @xcite models might predict .
likewise , that quasar masses do not trace the red galaxy @xmath42 , as they would if quasar activity was long - lived or randomly ( but uniformly as a function of mass ) episodic in high - mass black holes . the former case would be expected from the low - level agn activity invoked in e.g. @xcite , if `` radio '' and optical or high - eddington ratio quasar activity were associated , but they are in fact generally believed to be distinct ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the latter case implies strong limits on implementations of e.g. the @xcite model , which seek to suppress cooling flows through sporadic but potentially high accretion rate agn activity .
although we can rule out some alternatives to the merger scenario , there remain a number of viable variants of `` quenching '' models , in which crossing the critical halo mass @xmath163 and entering a `` hot '' accretion regime plays a key role in the `` transition '' to the rs .
especially given that some feedback mechanism is typically required , even in the `` hot '' accretion regime , to prevent the formation of cooling flows , it is easy to imagine a scenario in which , upon entering this regime , new infalling halo gas is shock - heated , but feedback from e.g. a central disk galaxy with a small black hole is inefficient , cold gas reservoirs remain large , and cooling flows can form . thus the system will not redden until it subsequently undergoes a major merger , which morphologically transforms the system , rapidly exhausts the remaining cold gas reservoir , and triggers a quasar and builds up a massive black hole , injecting some level of feedback and enabling efficient future ( e.g. cyclic agn or radio - mode ) feedback . the `` hot '' accretion regime may be a necessary prerequisite for feedback to efficiently prevent subsequent cooling , and as discussed in [ sec : halomass ] , our comparisons are all consistent with this possibility .
this is generally similar to the scenario assumed in e.g. @xcite , although they do explicitly incorporate quasar light curves or feedback . recognizing these distinctions ( as opposed to e.g. assuming a system simply `` shuts down '' upon reaching @xmath163 )
will probably have little effect on the @xmath8 predictions of semi - analytic models , since in either case star formation will be effectively suppressed at relatively early times in the most massive systems ( see also * ? ? ?
. however , at higher redshifts when massive objects are still forming , the distinctions will almost certainly be significant .
we note that none of our conclusions conflict with the hypothesis that , once formed , elliptical galaxies can continue to grow by dry mergers . however , they emphasize that the importance of such mergers is restricted to the most massive galaxies at low redshifts . our results ,
even the steep evolution of @xmath21 implying some `` downsizing '' in red galaxy assembly , are all consistent with ( and in fact , marginally _ favor _ ) the relatively low observationally inferred dry merger rate ( @xmath100 major dry merger since @xmath5 ) . essentially , `` downsizing '' in galaxy assembly as we have quantified it is not , strictly speaking , `` anti - hierarchical . ''
massive galaxies still continue to build up their populations to the present ; it is simply a statement that the _ relative _ rates of red galaxy formation / assembly decrease or `` slow down '' at late times in the most massive systems .
this could be related to pure dark matter processes , for example the rapid evolution in large overdensities could simply exhaust the `` supply '' of galaxies with which to merge , or the cluster environments of massive systems at low redshift attain sufficient circular velocities as to rapidly reduce merger rates ( see also * ? ? ?
the evolution of the `` transition '' mass may , alternatively , be a statement that galaxy assembly does not strictly trace halo assembly .
there are a number of baryonic processes which make this possible , as it simply requires that the effective baryon conversion efficiencies in galaxies be a function of time , or different for central vs. satellite systems .
improved measurements of early - type mass functions at high redshift ( @xmath5 ) , larger samples of mergers from which to construct merger mass functions , revised or direct determinations of high - redshift conditional mass functions , and direct observations of the masses of quasar hosts will substantially improve the constraints in this paper .
ultimately , the integration of the merger and quasar host mass functions may enable a purely observational comparison with the remnant , red galaxy mass function .
calibration of the observable merger `` timescale '' with realistic high resolution galaxy merger simulations , i.e. calibration of selection efficiencies for observed merger fractions , can further remove the factor @xmath40 uncertainty in comparing the rates of elliptical buildup and observed merger populations .
the association favored here between mergers , quasars , and elliptical buildup also makes specific predictions for the characteristic masses of e+a galaxies and quasar hosts as a function of redshift , which should be testable in future wide - field surveys .
the scenario we have described does not , of course , imply that mergers , quasars , and remnant ellipticals will necessarily be recognizable as the same , singular objects at a given instant in fact , simulations which follow the transition through these stages ( e.g. , * ? ? ?
* ) , predict that they will be seen as distinct phases in merger - triggered evolution , and observations tracking e.g. the associations between dynamical merger state and quasar activity ( e.g. , * ? ? ?
* ) support this distinction .
what we ultimately find evidence for here in the masses , luminosities , and clustering properties of mergers , galaxies being `` added '' or `` in transition '' to the red sequence , and quasars is that they are drawn from the same `` parent '' population , and that this population is distinct from the `` quiescent '' all / red / blue galaxy population . again , none of this strictly implies causality , but it does favor models which associate these populations with the same event , a natural expectation if mergers of gas - rich galaxies trigger quasars and morphologically transform disks to spheroids , moving `` new '' mass to the early - type population and leaving an elliptical , gas - poor , rapidly reddening remnant galaxy .
we thank eric bell , arjun dey , david hogg , casey papovich , rachel somerville , stijn wuyts , sandy faber , and marijn franx for very helpful discussions .
we also thank t. j. cox , brant robertson , and the anonymous referee whose comments improved this manuscript .
this work was supported in part by nsf grants aci 96 - 19019 , ast 00 - 71019 , ast 02 - 06299 , and ast 03 - 07690 , and nasa atp grants nag5 - 12140 , nag5 - 13292 , and nag5 - 13381 .
adelberger , k. l. , & steidel , c. c. 2005a , , 627 , l1 adelberger , k. l. , & steidel , c. c. 2005b , , 630 , 50 alexander , d. m. , et al . 2005a , nature , 434 , 738 alexander , d. m. , bauer , f. e. , chapman , s. c. , smail , i. , blain , a. w. , brandt , w. n. , & ivison , r. j. 2005b , , 632 , 736 balogh , m. l. , et al .
2004 , , 615 , l101 barnes , j. e. & hernquist , l. 1991 , apj , 370 , l65 barnes , j. e. & hernquist , l. 1992 , ara&a , 30 , 705 barnes , j. e. & hernquist , l. 1996 , apj , 471 , 115 baugh , c. m. , lacey , c. g. , frenk , c. s. , granato , g. l. , silva , l. , bressan , a. , benson , a. j. , & cole , s. 2005 , , 356 , 1191 bell , e. f. , mcintosh , d. h. , katz , n. , & weinberg , m. d. 2003 , , 149 , 289 bell , e. f. , et al . 2004 , , 608 , 752 bell , e. f. , et al . 2006a , , 640 , 241 bell , e. f. , phleps , s. , somerville , r. s. , wolf , c. , borch , a. , & meisenheimer , k. 2006b , , 652 , 270 benson , a. j. , bower , r. g. , frenk , c. s. , lacey , c. g. , baugh , c. m. , & cole , s. 2003 , , 599 , 38 berrier , j. c. , bullock , j. s. , barton , e. j. , guenther , h. d. , zentner , a. r. , & wechsler , r. h. 2006 , , 652 , 56 binney , j. 2004 , , 347 , 1093 birnboim , y. , & dekel , a. 2003 , , 345 , 349 blake , c. , et al .
2004 , , 355 , 713 blumenthal , g. r. , faber , s. m. , primack , j. r. , & rees , m. j. 1984 , , 311 , 517 bond , j. r. , cole , s. , efstathiou , g. , & kaiser , n. 1991 , , 379 , 440 borch , a. , et al .
2006 , , 453 , 869 borys , c. , smail , i. , chapman , s. c. , blain , a. w. , alexander , d. m. , & ivison , r. j. 2005 , , 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 , , 370 , 645 boyle , b. j. , shanks , t. , croom , s. m. , smith , r. j. , miller , l. , loaring , n. , & heymans , c. 2000 , , 317 , 1014 brinchmann , j. , et al .
1998 , , 499 , 112 brinchmann , j. , & ellis , r. s. 2000 , , 536 , l77 brown , m. j. i. , jannuzi , b. t. , dey , a. , & tiede , g. p. 2005
, , 621 , 41 brown , m. j. i. , dey , a. , jannuzi , b. t. , brand , k. , benson , a. j. , brodwin , m. , croton , d. j. , & eisenhardt , p. r. 2006 , , in press [ astro - ph/0609584 ] bruzual , g. , & charlot , s. 2003 , , 344 , 1000 bundy , k. , fukugita , m. , ellis , r. s. , kodama , t. , & conselice , c. j. 2004 , , 601 , l123 bundy , k. , ellis , r. s. , & conselice , c. j. 2005a , , 625 , 621 bundy , k. , et al .
2006 , , 651 , 120 caldwell , n. , rose , j. a. , franx , m. , & leonardi , a. j. 1996 , , 111 , 78 caldwell , n. , rose , j. a. , & concannon , k. d. 2003 , , 125 , 2891 cassata , p. , et al .
2005 , , 357 , 903 cattaneo , a. , dekel , a. , devriendt , j. , guiderdoni , b. , & blaizot , j. 2006 , , 370 , 1651 cimatti , a. , daddi , e. , & renzini , a. 2006 , , 453 , l29 cole , s. , et al .
2001 , , 326 , 255 conselice , c. j. , bershady , m. a. , dickinson , m. , & papovich , c. 2003 , , 126 , 1183 conselice , c. j. , blackburne , j. a. , & papovich , c. 2005 , , 620 , 564 cooray , a. 2005 , , 364 , 303 cooray , a. 2006 , , 365 , 842 couch , w. j. , & sharples , r. m. 1987 , , 229 , 423 cowie , l. l. , songaila , a. , hu , e. m. , & cohen , j. g. 1996 , , 112 , 839 cox , t. j. , di matteo , t. , hernquist , l. , hopkins , p. f. , robertson , b. , & springel , v. 2006a , , 643 , 692 cox , t. j. , dutta , s. n. , di matteo , t. , hernquist , l. , hopkins , p. f. , robertson , b. , & springel , v. 2006b , , 650 , 791 croom , s. m. , shanks , t. , boyle , b. j. , smith , r. j. , miller , l. , loaring , n. s. , & hoyle , f. 2001 , , 325 , 483 croom , s. m. , smith , r. j. , boyle , b. j. , shanks , t. , miller , l. , outram , p. j. , & loaring , n. s. 2004 , , 349 , 1397 croom , s. m. , et al .
2005 , , 356 , 415 croton , d. j. , et al .
2006 , , 365 , 11 dasyra , k. m. , et al .
2006 , , 638 , 745 dekel , a. , & birnboim , y. 2006 , , 368 , 2 de lucia , g. , & blaizot , j. 2006 , , in press [ astro - ph/0606519 ] de propris , r. , liske , j. , driver , s. p. , allen , p. d. , & cross , n. j. g. 2005 , , 130 , 1516 di matteo , t. , springel , v. , & hernquist , l. 2005 , nature , 433 , 604 di serego alighieri , s. , et al .
2005 , , 442 , 125 dressler , a. , & gunn , j. e. 1983 , , 270 , 7 driver , s. p. , liske , j. , cross , n. j. g. , de propris , r. , & allen , p. d. 2005 , , 360 , 81 drory , n. , bender , r. , feulner , g. , hopp , u. , maraston , c. , snigula , j. , & hill , g. j. 2004 , , 608 , 742 eisenstein , d. j. , & hu , w. 1999 , , 511 , 5 elvis , m. , et al .
1994 , , 95 , 1 faber , s. m. , et al .
2005 , , in press [ astro - ph/0506044 ] farrah , d. , et al . 2006 , , 641 , l17 farrah , d. , geach , j. , fox , m. , serjeant , s. , oliver , s. , verma , a. , kaviani , a. , & rowan - robinson , m. 2004 , , 349 , 518 ferrarese , l. & merritt , d. 2000 , , 539 , l9 feulner , g. , bender , r. , drory , n. , hopp , u. , snigula , j. , & hill , g. j. 2003 , , 342 , 605 floyd , d. j. e. , kukula , m. j. , dunlop , j. s. , mclure , r. j. , miller , l. , percival , w. j. , baum , s. a. , & odea , c. p. 2004
, , 355 , 196 fontana , a. , et al .
2004 , , 424 , 23 fontana , a. , et al .
2006 , , 459 , 745 franceschini , a. , et al .
2006 , , 453 , 397 franx , m. 1993 , , 407 , l5 gallazzi , a. , charlot , s. , brinchmann , j. , & white , s. d. m. 2006 , , 370 , 1106 gao , l. , springel , v. , & white , s. d. m. 2005 , , 363 , l66 gebhardt , k. , bender , r. , bower , g. et al .
2000 , , 539 , l13 goto , t. 2005 , , 357 , 937 graham , a. w. , erwin , p. , caon , n. , & trujillo , i. 2001 , , 563 , l11 granato , g. l. , de zotti , g. , silva , l. , bressan , a. , & danese , l. 2004 , , 600 , 580 haiman , z. , & hui , l. 2001 , , 547 , 27 hasinger , g. , miyaji , t. , & schmidt , m. 2005 , , 441 , 417 hennawi , j. f. , et al . 2006 , , 131 , 1 ho , l. c. 2002 , , 564 , 120 hopkins , p. f. , hernquist , l. , martini , p. , cox , t. j. , robertson , b. , di matteo , t. , & springel , v. 2005a , , 625 , l71 hopkins , p. f. , hernquist , l. , cox , t. j. , robertson , b. , di matteo , t. , martini , p. , & springel , v. 2005b , , 630 , 705 hopkins , p. f. , hernquist , l. , cox , t. j. , robertson , b. , di matteo , t. , & springel , v. 2005c , , 630 , 716 hopkins , p. f. , hernquist , l. , cox , t. j. , robertson , b. , di matteo , t. , & springel , v. 2005d , , 632 , 81 hopkins , p. f. , hernquist , l. , cox , t. j. , robertson , b. , di matteo , t. , & springel , v. 2006a , , 639 , 700 hopkins , p. f. , hernquist , l. , cox , t. j. , robertson , b. , di matteo , t. , & springel , v. 2006b , , 163 , 50 hopkins , p. f. , hernquist , l. , cox , t. j. , robertson , b. , & springel , v. 2006c , , 163 , 50 hopkins , p. f. , somerville , r. s. , hernquist , l. , cox , t. j. , robertson , b. , & li , y. 2006 , , 652 , 864 hopkins , p. f. , & hernquist , l. 2006 , , 166 , 1 hopkins , p. f. , narayan , r. , & hernquist , l. 2006 , , 643 , 641 hopkins , p. f. , richards , g. t. , & hernquist , l. 2007 , , 654 , 731 juneau , s. , et al .
2005 , , 619 , l135 kauffmann , g. & haehnelt , m. g. 2000 , , 311 , 576 kauffmann , g. , & haehnelt , m. g. 2002 , , 332 , 529 kauffmann , g. , et al .
2003 , , 341 , 54 kelson , d. d. , illingworth , g. d. , van dokkum , p. g. , & franx , m. 2000 , , 531 , 184 kere , d. , katz , n. , weinberg , d. h. , & dav , r. 2005 , , 363 , 2 krding , e. g. , jester , s. , & fender , r. 2006 , , 372 , 1366 kollmeier , j. a. , et al .
2006 , , 648 , 128 lapi , a. , shankar , f. , mao , j. , granato , g. l. , silva , l. , de zotti , g. , & danese , l. 2006 , , 650 , 42 le fvre , o. , et al .
2000 , , 311 , 565 lemson , g. , & kauffmann , g. 1999 , , 302 , 111 li , c. , kauffmann , g. , jing , y. p. , white , s. d. m. , brner , g. , & cheng , f. z. 2006 , , 368 , 21 lidz , a. , hopkins , p. f. , cox , t. j. , hernquist , l. , & robertson , b. 2006 , , 641 , 41 lin , l. , et al .
2004 , , 617 , l9 lotz , j. m. , primack , j. , & madau , p. 2004
, , 128 , 163 lotz , j. m. , madau , p. , giavalisco , m. , primack , j. , & ferguson , h. c. 2006a , , 636 , 592 lotz , j. m. , et al . 2006b , , in press [ astro - ph/0602088 ] madgwick , d. , et al .
2003 , , 599 , 997 magorrian , j. et al .
1998 , , 115 , 2285 mandelbaum , r. , seljak , u. , kauffmann , g. , hirata , c. m. , & brinkmann , j. 2006 , , 368 , 715 marconi , a. , et al .
2004 , , 351 , 169 marconi , a. , & hunt , l.k .
2003 , , 589 , l21 martini , p. 2004 , in carnegie obs .
1 , coevolution of black holes and galaxies , ed . l.c .
ho ( cambridge : cambridge univ . press ) , 170 martini , p. , & weinberg , d. h. 2001 , , 547 , 12 masjedi , m. , et al .
2006 , , 644 , 54 mclure , r. j. , & dunlop , j. s. 2004 , , 352 , 1390 menci , n. , fontana , a. , giallongo , e. , grazian , a. , & salimbeni , s. 2006 , , 647 , 753 meneux , b. , et al .
2006 , , 452 , 387 mihos , j. c. & hernquist , l. 1994 , apj , 431 , l9 mihos , j. c. & hernquist , l. 1996 , apj , 464 , 641 mo , h. j. , & white , s. d. m. 1996 , , 282 , 347 monaco , p. , & fontanot , f. 2005 , , 359 , 283 myers , a. d. , et al .
2006 , , 638 , 622 neistein , e. , van den bosch , f. c. , & dekel , a. 2006 , , 372 , 933 nelan , j. e. , et al .
2005 , , 632 , 137 norberg , p. , et al .
2002 , , 332 , 827 norton , s. a. , gebhardt , k. , zabludoff , a. i. , & zaritsky , d. 2001 , , 557 , 150 pannella , m. , hopp , u. , saglia , r. p. , bender , r. , drory , n. , salvato , m. , gabasch , a. , & feulner , g. 2006 , , 639 , l1 papovich , c. , et al .
2006 , , 640 , 92 patton , d. r. , et al .
2002 , , 565 , 208 peng , c. y. , impey , c. d. , rix , h .- w . ,
kochanek , c. s. , keeton , c. r. , falco , e. e. , lehr , j. , & mcleod , b. a. 2006 , , 649 , 616 percival , w. j. , scott , d. , peacock , j. a. , & dunlop , j. s. 2003 , , 338 , l31 pollo , a. , et al . 2006 , , 451 , 409 pozzetti , l. , et al .
2003 , , 402 , 837 quintero , a. d. , et al .
2004 , , 602 , 190 rees , m. j. , & ostriker , j. p. 1977
, , 179 , 541 renzini , a. 2006 , , 44 , 141 richards , g. t. , et al .
2005 , , 360 , 839 richards , g. t. , et al .
2006 , , 131 , 2766 richards , g. t. , et al .
2006 , , 166 , 470 robertson , b. , hernquist , l. , cox , t. j. , di matteo , t. , hopkins , p. f. , martini , p. , & springel , v. 2005a , , 641 , 90 robertson , b. , cox , t. j. , hernquist , l. , franx , m. , hopkins , p. f. , martini , p. , & springel , v. 2005b , , 641 , 21 rothberg , b. , & joseph , r. d. 2004 , , 128 , 2098 rothberg , b. , & joseph , r. d. 2006 , , 131 , 185 salpeter , e. e. 1955 , , 121 , 161 sanders , d. b. et al .
1988 , , 328 , l35 scannapieco , e. , & oh , s. p. 2004
, , 608 , 62 scannapieco , e. , silk , j. , & bouwens , r. 2005 , , 635 , l13 schweizer , f. 1996 , , 111 , 109 serber , w. , bahcall , n. , mnard , b. , & richards , g. 2006 , , 643 , 68 shankar , f. , lapi , a. , salucci , p. , de zotti , g. , & danese , l. 2006 , , 643 , 14 shepherd , c. w. , carlberg , r. g. , yee , h. k. c. , morris , s. l. , lin , h. , sawicki , m. , hall , p. b. , & patton , d. r. 2001 , , 560 , 72 sheth , r. k. , mo , h. j. , & tormen , g. 2001 , , 323 , 1 shields , g. a. , et al .
2003 , , 583 , 124 silk , j. & rees , m. j. 1998 , , 331 , l1 somerville , r. s. , primack , j. r. , & faber , s. m. 2001 , , 320 , 504 somerville , r. s. , et al . 2004a , , 600 , l135 somerville , r. s. , lee , k. , ferguson , h. c. , gardner , j. p. , moustakas , l. a. , & giavalisco , m. 2004b , , 600 , l171 springel , v. , di matteo , t. , & hernquist , l. 2005 , , 620 , l79 springel , v. , di matteo , t. , & hernquist , l. 2005 , , 361 , 776 strateva , i. , et al .
2001 , aj , 122 , 1861 straughn , a. n. , cohen , s. h. , ryan , r. e. , hathi , n. p. , windhorst , r. a. , & jansen , r. a. 2006 , , 639 , 724 tacconi , l. j. , genzel , r. , lutz , d. , rigopoulou , d. , baker , a. j. , iserlohe , c. , & tecza , m. 2002 , , 580 , 73 toledo , h. m. h. , dultzin - hacyan , d. , gonzalez , j. j. , & sulentic , j. w. 1999 , , 118 , 108 a. , 1977 , in evolution of galaxies and stellar populations , 401 , yale univ .
obs : new haven tran , k .- v .
h. , franx , m. , illingworth , g. , kelson , d. d. , & van dokkum , p. 2003
, , 599 , 865 tran , k .- v . h. , franx , m. , illingworth , g. d. , van dokkum , p. , kelson , d. d. , & magee , d. 2004 , , 609 , 683 treu , t. , ellis , r. s. , liao , t. x. , & van dokkum , p. g. 2005 , , 622 , l5 ueda , y. , akiyama , m. , ohta , k. , & miyaji , t. 2003 , , 598 , 886 vanden berk , d. e. , et al . 2006 , , 131 , 84 van der wel , a. , franx , m. , van dokkum , p. g. , & rix , h .- w . 2004 , , 601 , l5 van der wel , a. , franx , m. , van dokkum , p. g. , rix , h .- w . , illingworth , g. d. , & rosati , p. 2005
, , 631 , 145 van dokkum , p. g. , & stanford , s. a. 2003 , , 585 , 78 van dokkum , p. g. 2005 , , 130 , 2647 vignali , c. , brandt , w. n. , & schneider , d. p. 2003
, , 125 , 433 volonteri , m. , haardt , f. , & madau , p. 2003
, , 582 , 559 volonteri , m. , salvaterra , r. , & haardt , f. 2006 , , 373 , 121 white , r. l. , helfand , d. j. , becker , r. h. , glikman , e. , & devries , w. 2006 , , in press [ astro - ph/0607335 ] willmer , c. n. a. , et al . 2006 , , 647 , 853 wolf , c. , wisotzki , l. , borch , a. , dye , s. , kleinheinrich , m. , & meisenheimer , k. 2003 , , 408 , 499 wolf , c. , et al .
2005 , , 630 , 771 wyithe , j. s. b. , & loeb , a. 2003 , , 595 , 614 xu , c. k. , sun , y. c. , & he , x. t. 2004 , , 603 , l73 yamada , t. , et al .
2005 , , 634 , 861 yan , r. , madgwick , d. s. , & white , m. 2003 , , 598 , 848 yang , x. , mo , h. j. , & van den bosch , f. c. 2003 , , 339 , 1057 yang , x. , mo , h. j. , jing , y. p. , & van den bosch , f. c. 2005 , , 358 , 217 zabludoff , a. i. , zaritsky , d. , lin , h. , tucker , d. , hashimoto , y. , shectman , s. a. , oemler , a. , & kirshner , r. p. 1996
, , 466 , 104 zheng , z. , et al .
2005 , , 633 , 791 zucca , e. , et al .
2006 , , 455 , 879 | we compile a number of recent observations to estimate the time - averaged rate of formation or `` buildup '' of red sequence galaxies , as a function of mass and redshift . comparing this with the mass functions of mergers and quasar hosts , and independently comparing their clustering properties as a function of redshift
, we find that these populations trace the same mass distribution , with the same characteristic masses and evolution , in the redshift interval @xmath0 .
knowing one of the quasar , merger , or elliptical mass / luminosity functions , it is possible to predict the others . allowing for greater model dependence
, we compare the rate of early - type `` buildup '' with the implied merger and quasar `` triggering '' rates as a function of mass and redshift , and find agreement .
we show that over this redshift range , observed gas - rich merger fractions can account for the entire bright quasar luminosity function , and buildup of the red sequence at all but the highest masses at low redshift ( @xmath1 at @xmath2 ) where dissipationless `` dry '' mergers appear to dominate .
this supports a necessary prediction of theories which postulate that mergers between gas - rich blue galaxies produce ellipticals with an associated phase of bright quasar activity , after which the remnant becomes red .
all of these populations , regardless of sample selection , trace a similar characteristic `` transition '' mass reflecting the characteristic mass above which the elliptical population is mostly ( @xmath3 ) assembled at a given redshift , which increases with redshift over the observed range in a manner consistent with previous suggestions that `` cosmic downsizing '' may apply to red galaxy _ assembly _ as well as star formation .
we show that these mass distributions as a function of redshift do not uniformly trace the all / red / blue galaxy population , ruling out models in which quasar activity is generically associated with either star formation or is long - lived in `` old '' systems . |
an _ ordered tree _ is a rooted tree in which children of each vertex are ordered . figure [ fig : ord ] shows all the ordered tree with @xmath4 vertices .
it is well known ( see ( * ? ? ?
* exercise 6.19 ) ) that the number of ordered trees with @xmath5 vertices is given by the @xmath6th catalan number @xmath7 .
7.00mm7.00mm123.00mm43.00 mm ( 328.82 , 102.05)(0,0 ) ( 0,0 ) vertices , title="fig : " ] an _ ordered labeled tree _ is an ordered tree whose vertices are labeled by distinct nonnegative integers . in most cases , an ordered labeled tree with @xmath5 vertices is identified with an ordered tree on the vertex set @xmath8:=\{0,\dots , n\}$ ] .
let @xmath0 be the set of ordered labeled trees on @xmath8 $ ] .
clearly the cardinality of @xmath0 is given by @xmath9 is a rising factorial . for a given ordered labeled tree @xmath10 , a _
maximal decreasing subtree _ of @xmath10 is defined by the maximal ordered subtree from the root with all edges being decreasing , denoted by @xmath11 .
figure [ fig : tree ] illustrates the maximal decreasing subtree of a given tree @xmath10 . let @xmath2 be the set of ordered labeled trees on @xmath8 $ ] with its maximal decreasing subtree having @xmath12 edges .
17.00mm16.14mm126.00mm54.14 mm ( 308.98 , 107.72)(0,0 ) ( 0,0),title="fig : " ] ( 42.52,64.81)7 ( 14.17,8.12)8 ( 99.21,8.12)0 ( 127.56,36.46)10 ( 70.87,8.12)4 ( 155.91,8.12)5 ( 99.21,64.81)6 ( 70.87,36.46)2 ( 73.70,93.16)9 ( 42.52,36.46)1 ( 70.87,64.81)3 ( 8.50,93.16)@xmath10 ( 240.94,64.81)7 ( 297.64,64.81)6 ( 269.29,36.46)2 ( 272.13,93.16)9 ( 240.94,36.46)1 ( 269.29,64.81)3 ( 206.93,93.16)@xmath11 in this paper we present a formula for @xmath13 , which makes a refined enumeration of @xmath0 , or a generalization of equation .
note that a similar refinement for the rooted ( unordered ) labeled trees was done before ( see @xcite ) , but the ordered case is more complicated and has quite different features .
from now on we will consider labeled trees only .
so we will omit the word labeled " . recall that @xmath2 is the set of ordered trees on @xmath8 $ ] with its maximal decreasing ordered subtree having @xmath12 edges .
let @xmath14 be the set of ordered trees on @xmath8 $ ] attached additional @xmath15 increasing leaves to decreasing tree with @xmath12 edges .
note that the set @xmath14 first appeared in the ph.d .
thesis @xcite of drake .
let @xmath16 be the set of forests on @xmath17:=\{1,2,\ldots , n\}$ ] consisting of @xmath12 ordered trees , where the @xmath12 roots are not ordered . in figure
[ fig : forest ] , the first two forests are the same , but the third one is a different forest in @xmath18 .
-3.00mm-3.86mm106.00mm14.14 mm ( 308.98 , 51.02)(0,0 ) ( 0,0),title="fig : " ] ( 45.35,8.12 ) ( 187.09,8.12 ) ( 255.12,8.12 ) ( 76.54,36.46 ) ( 133.23,36.46 ) ( 246.61,36.46 ) ( 34.02,36.46 ) ( 175.75,36.46 ) ( 289.13,36.46 ) ( 17.01,8.12 ) ( 158.74,8.12 ) ( 300.47,8.12 ) ( 93.54,22.29 ) ( 206.93,22.29 ) define the numbers @xmath19 we will show that an ordered tree can be decomposed " into an ordered tree in @xmath20 and a forest in @xmath21 .
thus it is crucial to count the numbers @xmath22 and @xmath23 .
[ lem : znk ] the numbers @xmath22 satisfy the recursion : @xmath24 with the following boundary conditions : @xmath25 where @xmath26 is defined by @xmath27 .
consider a tree @xmath28 in @xmath14 .
the tree @xmath28 with @xmath5 vertices consists of its maximal decreasing tree with @xmath3 vertices and the number of increasing leaves is @xmath29 .
note that the vertex @xmath30 is always contained in @xmath31 .
if the vertex @xmath30 is a leaf of @xmath28 , consider the tree @xmath32 by deleting the leaf @xmath30 from @xmath28 .
the number of vertices in @xmath32 and @xmath33 are @xmath6 and @xmath12 , respectively .
so the number of possible trees @xmath32 is @xmath34 .
since we can not attach the vertex @xmath30 to @xmath29 increasing leaves in recovering @xmath28 , there are @xmath35 ways of recovering @xmath28 . thus the number of @xmath28 with the leaf @xmath30 is @xmath36
if the vertex @xmath30 is not a leaf of @xmath28 , then the vertex @xmath30 has at least one increasing leaf .
let the vertex @xmath37 be the leftmost leaf of the vertex @xmath30 and consider the tree @xmath38 obtained by deleting the leaf @xmath37 from @xmath28 .
the number of vertices in @xmath38 and @xmath39 are @xmath6 and @xmath3 , respectively .
so the number of possible trees @xmath38 is @xmath40 .
to recover @xmath28 is to relabel @xmath38 with @xmath8 \setminus { \left\{\ell\right\}}$ ] and to attach the vertex @xmath37 to the vertex @xmath30 . since the number
@xmath37 may be the number from @xmath41 to @xmath6 , the number of @xmath28 without the leaf @xmath30 is @xmath42 which completes the proof of recursion .
since @xmath43 is the set of decreasing ordered trees on @xmath8 $ ] , the equation holds @xcite with the convention @xmath44 . for @xmath45 or @xmath46
, @xmath14 should be empty , so the equation also holds .
[ lem : fnk ] for @xmath47 , we have @xmath48 with @xmath49 consider a forest @xmath50 in @xmath16 .
the forest @xmath50 consists of ( non - ordered ) @xmath12 ordered trees @xmath51 with roots @xmath52 , where @xmath53 . the number of ways for choosing roots @xmath54 from @xmath17 $ ] is equal to @xmath55 . from the _ reverse prfer algorithm ( rp algorithm )
_ in @xcite , the number of ways for adding @xmath29 vertices successively to @xmath12 roots @xmath54 is equal to @xmath56 for @xmath57 , thus the equation holds . by definition , @xmath58 is the set of the empty forest .
so @xmath59 .
since the number @xmath22 is determined by the recurrence relation in lemma [ lem : znk ] , we can count the number @xmath60 with the following theorem .
[ thm : main ] we have @xmath61 and @xmath62 , where @xmath63 is a rising factorial . given an ordered tree @xmath10 in @xmath2 ,
let @xmath28 be the subtree of @xmath10 consisting of @xmath11 and its increasing edges .
if the number of vertices of @xmath28 is @xmath64 , then @xmath28 is a subtree of @xmath10 with @xmath65 increasing leaves . also , the induced subgraph @xmath66 of @xmath10 generated by the @xmath15 vertices not belonging to @xmath11 is a ( non - ordered ) forest consisting of @xmath65 ordered trees whose roots are only increasing leaves of @xmath28
. now let us count the number of ordered trees @xmath67 with @xmath68 where @xmath69 is the set of vertices in @xmath28 .
first of all , the number of ways for selecting a set @xmath70 $ ] is equal to @xmath71 . by attaching @xmath65 increasing leaves to a decreasing tree with @xmath12 edges
, we can make an ordered trees on @xmath69 .
there are exactly @xmath72 ways for making such an ordered subtree on @xmath69 . by the definition of @xmath16 and lemma [ lem : fnk ] ,
the number of ways for constructing the other parts on @xmath73 is equal to @xmath74 since the range of @xmath75 is @xmath76 , the equation holds .
finally , @xmath77 is the set of decreasing ordered trees on @xmath8 $ ] , so @xmath78 holds for @xmath79 .
due to theorem [ thm : main ] , we can calculate @xmath60 for all @xmath6 , @xmath12 .
however a closed form , a recurrence relation , or a generating function of @xmath60 have not been found yet . the following might be a direction for solving the problem : shor @xcite showed that the number @xmath80 , which is the number of rooted trees on @xmath17 $ ] with @xmath12 improper edges , satisfies @xmath81 where an edge @xmath82 is called _ improper _ if @xmath83 is the endpoint closer to root and @xmath83 has a larger label than some descendant of @xmath84 .
zeng @xcite found that the generating function for @xmath85 is the ramanujan polynomial @xmath86 , which is defined by @xmath87 drake @xcite observed that @xmath88 for all @xmath89 , by using the generating function method .
actually , @xmath22 and @xmath90 satisfy the same recursion and initial conditions , so we are able to construct a recursive bijection between these two objects . with this point of view
, it would be interesting to find a certain set of rooted trees of cardinality @xmath60 .
this research was supported by basic science research program through the national research foundation of korea ( nrf ) funded by the ministry of education , science and technology ( 2011 - 0008683 , 2012r1a1a1014154 ) . | let @xmath0 be the set of ordered labeled trees on @xmath1 .
a maximal decreasing subtree of an ordered labeled tree is defined by the maximal ordered subtree from the root with all edges being decreasing . in this paper
, we study a new refinement @xmath2 of @xmath0 , which is the set of ordered labeled trees whose maximal decreasing subtree has @xmath3 vertices . |
this paper is a continuation of studies initiated in @xcite regarding infinite families of non - isotopic and symplectic tori representing the same homology class in a symplectic 4-manifold .
let @xmath6 be the rational elliptic surface obtained by blowing up at nine branch points of a generic pencil of cubic curves in @xmath8 .
let @xmath0 denote the complex surface obtained by performing a logarithmic transformation of multiplicity @xmath2 on a regular fiber @xmath9 of @xmath10 .
our main result is the following : [ theorem : main ] for any integer @xmath5 , there exists an infinite family of pairwise non - isotopic symplectic tori representing the primitive homology class @xmath4 \in h_2(e(1)_p)$ ] , where @xmath1 is the multiple fiber of @xmath0 .
even though the tori we construct are not only non - isotopic but also inequivalent in the sense that there is no self - diffeomorphism of the ambient 4-manifold that carries one torus to another , we state the theorem the way we do since our interest in the subject is mainly due to the more general problem of finding non - isotopic symplectic representatives of a homology class , known as the symplectic isotopy problem . it is well - known that there exists a diffeomorphism between @xmath0 and @xmath10 , sending the homology class @xmath4 $ ] to @xmath11\,$ ] ( see e.g.@xmath12@xcite ) . hence we immediately obtain the following : [ cor : main ] for a suitable choice of a symplectic form on @xmath10 , there exists an infinite family of pairwise non - isotopic symplectic tori representing the primitive homology class @xmath11 \in h_2(e(1))$ ] , where @xmath9 is the the torus fiber of the elliptic fibration @xmath13 .
examples of homologous , non - isotopic , symplectic tori were first constructed in @xcite and then in @xcite , @xcite and @xcite .
infinite families of non - isotopic symplectic tori representing divisible classes @xmath14\in h_2(e(1)_p)$ ] , @xmath15 , were constructed in @xcite . in @xcite and @xcite , symplectic tori representing the same primitive class in symplectic 4-manifolds homeomorphic but not diffeomorphic to @xmath10 were constructed .
it should be noted that the non - existence of such an infinite family of tori in the rational surfaces @xmath8 and @xmath16 is proved by sikorav in @xcite and by siebert and tian in @xcite , respectively .
moreover , in contrast with corollary [ cor : main ] , which easily generalizes to @xmath17 for @xmath18 , it is conjectured that there is at most one symplectic torus ( up to isotopy ) representing each homology class in @xmath17 for @xmath19 .
the proof of theorem [ theorem : main ] is spread out over the next four sections . after reviewing the relevant definitions we interpret @xmath0 as a link surgery manifold in section [ sec : link surgery ] and construct a family of homologous symplectic tori in @xmath0 in section [ sec : symplectic family ] . in section [ sec : alexander ] , alexander polynomials of certain braid closures are given and then in section [ sec : sw ] these polynomials are used to compute the seiberg - witten invariants of fiber sums of @xmath0 with @xmath10 , along the tori we construct and a regular fiber @xmath9 , respectively , which in turn give the non - isotopy of these tori . in section [ sec : generalization ] , we present a direct generalization in the form of theorem [ theorem : generalization ] . in this introduction and
elsewhere in the paper by isotopy we mean smooth isotopy and all homology groups have integer coefficients .
in this section , first we review the generalization of the link surgery construction of fintushel and stern @xcite by vidussi @xcite , and then give specific link surgeries that will be used in the following sections . for an @xmath20-component link @xmath21 ,
choose an ordered homology basis of simple closed curves @xmath22 such that the pair @xmath23 lie in the @xmath24-th boundary component of the link exterior and the intersection of @xmath25 and @xmath26 is 1 .
let @xmath27 ( @xmath28 ) be a 4-manifold containing a 2-dimensional torus submanifold @xmath29 of self - intersection @xmath30 .
choose a cartesian product decomposition @xmath31 , where each @xmath32 ( @xmath33 ) is an embedded circle in @xmath27 .
[ def : data ] the ordered collection @xmath34 is called a _ link surgery gluing data _ for an @xmath20-component link @xmath35 .
we define the _ link surgery manifold corresponding to _
@xmath36 to be the closed @xmath7-manifold @xmath37\hspace{-20pt}\bigcup_{f_i\times\partial d^2=(s^1\times \alpha_i)\times\beta_i}\hspace{-20pt } [ s^1\times(s^3\setminus \nu l)]\ , , \ ] ] where @xmath38 denotes the tubular neighborhoods . here
, the gluing diffeomorphisms between the boundary 3-tori identify the torus @xmath39 of @xmath27 with @xmath40 factor - wise , and act as the complex conjugation on the last remaining @xmath41 factor .
strictly speaking , the diffeomorphism type of the link surgery manifold @xmath42 may possibly depend on the chosen trivialization of @xmath43 ( the framing of @xmath29 ) .
however , we will suppress this dependence in our notation .
it is well known ( see e.g. @xcite ) that the diffeomorphism type of @xmath42 is independent of the framing of @xmath29 when @xmath44 .
we fix a cartesian product decomposition of a regular torus fiber @xmath45 in the elliptic surface @xmath10 . also fix @xmath46 . ]
let @xmath47 be the hopf link in figure @xmath48 .
for the link surgery gluing data @xmath49 we shall denote @xmath42 by @xmath50 . here ,
@xmath51 and @xmath52 denote the meridian and the longitude of a knot , respectively . in particular , when @xmath53 , we denote @xmath42 by @xmath0 .
this notation is consistent with the existing literature as there is an obvious diffeomorphism ( see diffeomorphism ( [ eq : cylinder ] ) in the proof of lemma [ lemma : main ] ) between our link surgery manifold @xmath54\,\cup\,[s^1\times(s^3\setminus l)]\,\cup\,[t^2\times d^2]\,\ ] ] and the logarithmic transform @xmath55 \cup_{\varphi } [ t^2\times d^2]$ ] , where the gluing diffeomorphism @xmath56 induces the linear map @xmath57 between the first homology groups with the obvious choice of bases .
[ lemma : e(1)_p is symplectic ] if @xmath58 , then @xmath0 is a symplectic @xmath7-manifold diffeomorphic to @xmath10 .
the diffeomorphism type of the logarithmic transform @xmath0 does not depend on the choice of the gluing map @xmath59 .
the homology class @xmath4\in h_2(e(1)_p)$ ] is primitive .
see theorem 3.3.3 , remark 3.3.5 and theorem 8.3.11 in @xcite .
let @xmath60 \subset e(1)_p$ ] , where the closed curve @xmath61 is given by figure [ fig:3d ] . note that @xmath62 could be isotoped to an @xmath63 torus knot in @xmath64 . in @xmath65 [ lemma
: main ] for every pair of integers @xmath5 and @xmath66 , @xmath67 is a symplectic submanifold of @xmath0 and we have @xmath68=[f_p]\,$ ] in @xmath69 . to determine the homology class of @xmath70 , note that @xmath71= m[\mu ( a ) ] + ( 1-mp)[\mu ( b)]$ ] in @xmath72 . when we glue @xmath73 to @xmath74 using the gluing data @xmath36 in ( [ eq : data ] ) , the homology class @xmath75-p[\mu(b)]=[\lambda ( b)]-p[\mu ( b)]$ ] gets identified with @xmath76 \in h_1(t^2\times d^2)$ ] , which is trivial . hence by knneth s theorem , @xmath68=[s^1\times \mu(b)]\,$ ] in @xmath77 .
finally , note that @xmath78=[f_p]\,$ ] and @xmath11=[s^1\times \mu(a)]=[s^1\times\lambda(b)]=[s^1 \times p\,\mu(b)]=p[f_p]\,$ ] in @xmath69 .
it is easy to see that the link exterior @xmath79 is diffeomorphic to @xmath80 , where @xmath81\,$ ] is an annulus .
hence we may identify the middle piece in the decomposition ( [ eq : decomposition ] ) above as the cylinder @xmath82 \:\cong\ : [ s^1 \times ( s^1 \times \mathbb{a } ) ] \:\cong\ : t^3\times [ 0,1 ] .\ ] ] since we assume that @xmath5 , the logarithmic transform corresponds to the rational blowdown construction of fintushel and stern @xcite . on the other hand , by @xcite , rational blowdown
can be done symplectically and the symplectic form needs to be modified only near the collar neighborhood of the boundary lens spaces involved . as a result of this and
since @xmath83 remains symplectic in @xmath0 , we may assume that the symplectic form on @xmath0 restricts to @xmath84 on the subset @xmath85\subset[s^1 \times ( s^1\times \mathbb{a})]$ ] for suitable @xmath86 , where @xmath87 and @xmath88 are the angular coordinates on the first and the second @xmath41 factors respectively , and @xmath89 are the polar coordinates on the annulus @xmath90 .
we can embed the curve @xmath62 inside @xmath91\subset(s^1\times \mathbb{a})$ ] such that @xmath62 is transverse to every annulus of the form , @xmath92 , and the restriction @xmath93 never vanishes .
it follows that @xmath94 , and consequently @xmath70 is a symplectic submanifold of @xmath0 .
[ lem : alex ] @xmath95 @xmath96 \ , , \ ] ] where @xmath97 is the 3-component link in figure [ fig:3d ] , and the variables @xmath87 , @xmath98 and @xmath99 correspond to the components @xmath100 , @xmath101 and @xmath62 respectively .
-strand braid whose closure gives @xmath102 this follows from the formula in theorem 1 of @xcite which gives the multivariable alexander polynomial of a braid closure and its axis in terms of the representation of the braid .
we view @xmath100 as the axis of the @xmath103-strand braid in figure [ fig : torusknotbraid ] .
note that the closure of this braid and its axis is @xmath104 .
this statement can also be thought of as a slightly modified special case of lemma 6.2 in @xcite .
[ rem : symalex ] note that the symmetrized alexander polynomial of @xmath97 is @xmath105
our strategy is to show that for fixed choice of @xmath2 the isotopy types of the tori @xmath106 can be distinguished by comparing the seiberg - witten invariants of the corresponding family of fiber sum 4-manifolds @xmath107 .
note that there is a canonical framing of a regular fiber @xmath9 in @xmath10 , coming from the elliptic fibration @xmath13 .
the fiber sum @xmath108 is diffeomorphic to the link surgery manifold @xmath109 , where @xmath110 & & \{(e(1),f = c_1\times c_2 ) , ( t^2\times s^2,f_2=s^1 \times s^1 \times\{{\rm pt } \}),(e(1 ) , f = c_1\times c_2)\}\big ) .
\nonumber\end{aligned}\ ] ] it was already observed in @xcite that the fiber sum construction corresponds to this type of link surgery .
( see also @xcite . )
recall that the seiberg - witten invariant @xmath111 of a 4-manifold @xmath112 can be thought of as an element of the group ring of @xmath113 , i.e. @xmath114 $ ] .
if we write @xmath115 , then we say that @xmath116 is a seiberg - witten _ basic class _ of @xmath112 if @xmath117 . since the seiberg - witten invariant of a 4-manifold is a diffeomorphism invariant ,
so are the divisibilities of seiberg - witten basic classes .
the seiberg - witten invariant of the link surgery manifold @xmath109 is known to be related to the alexander polynomial @xmath118 of the link @xmath97 .
[ theorem : seiberg - witten ] let @xmath119\rightarrow l_{m , p}(\mathfrak{d}')$ ] be the inclusion map .
let @xmath120,$ ] @xmath121 \in h_2(l_{m , p}(\mathfrak{d } ' ) ) .$ ] then @xmath122 and @xmath123 are linearly independent . the seiberg - witten invariant of @xmath109 is given by @xmath124 \nonumber \ , \ , .
\end{aligned}\ ] ] let @xmath125 , and let @xmath126 .
recall from @xcite that we have @xmath127 , and also @xmath128^{-1}-[t^2\times\{{\rm pt } \ } ] } \ ; .\ ] ] from the gluing formulas in @xcite and , we may conclude that @xmath129^{-1}-[t^2\times\{{\rm pt } \ } ] } \cdot \delta_{l_{m , p}}^{\rm sym}(\xi^2,\varsigma^2,\tau^2 ) \
, , \end{aligned}\ ] ] where @xmath130 $ ] .
first of all , @xmath131 = [ \mu ( a ) ] + ( 1-mp ) [ \mu ( c ) ] \in h_1(n)$ ] . on the other hand , according to the gluing data @xmath132 , @xmath133 is identified with @xmath134 , therefore @xmath131 = p [ \mu ( b)]$ ] in @xmath109 and consequently @xmath135 also note that @xmath136 and @xmath137 are identified by the gluing data @xmath132 , hence @xmath138
= \varsigma \in h_2 ( l_{m , p}(\mathfrak{d } ' ) ) \ , .\ ] ] from lemma [ lem : alex ] and remark [ rem : symalex ] we have @xmath139 \ , .
\nonumber\end{aligned}\ ] ] after putting equations ( [ eqn : prod ] ) , ( [ eqn : xi ] ) , ( [ eqn : t2 ] ) and ( [ eqn : symalex ] ) together and simplifying the resulting expression , we get ( [ eq : sw ] ) . next
we show that @xmath122 and @xmath123 are linearly independent elements of @xmath140 . a mayer - vietoris argument , combined with freedman s classification theorem ( see , for example , @xmath12@xcite )
, shows that @xmath109 is homeomorphic to @xmath141 .
consider the composition of homomorphisms @xmath142 where the first map is a part of the knneth isomorphism @xmath143 note that @xmath144 , as is easily seen from the long exact sequence of the pair @xmath145 as follows .
@xmath146 { \mathbb{z } } & \longrightarrow & { \mathbb{z}}^3 & \longrightarrow & h_2(n ) & \stackrel{0}{\longrightarrow } & { \mathbb{z}}^3 & \longrightarrow & { \mathbb{z}}^6 \end{array}\ ] ] note that the first map sends the generator @xmath147 to the diagonal element @xmath148 , while the last map is injective .
we have also used the lefschetz duality theorem ( for manifolds with boundary ) to identify @xmath149 .
next consider the long exact sequence of the pair @xmath150 : @xmath151 the kernel of the last map @xmath152 is isomorphic to @xmath153 . by lefschetz duality theorem ( for relative manifolds ) , @xmath153 is in turn isomorphic to @xmath154 since we have @xmath155 and @xmath156 the kernel of @xmath152 is isomorphic to @xmath157 .
finally we observe that only one @xmath158 summand of ( [ eq : z plus z ] ) lies in the kernel of the composition ( [ eq : homomorphism ] ) .
the other @xmath158 summand belongs to the kernel of @xmath159 where the first map is the second part of the knneth isomorphism ( [ eq : kunneth ] ) .
we have thus shown that the kernel of the composition ( [ eq : homomorphism ] ) is of rank one .
it follows immediately that @xmath122 and @xmath123 are linearly independent , since we have already showed that @xmath160 .
a more detailed analysis shows that @xmath161 can be extended to a basis of @xmath140 , which we shall omit .
( also see the proof of proposition 3.2 in @xcite for a similar argument . )
[ cor : basics ] the number of seiberg - witten basic classes of @xmath162 is @xmath163 first of all , the summands in equation([eq : sw ] ) have no common terms , i.e. terms with matching powers for @xmath122 and @xmath123 , because the general term in the first summand is @xmath164 , where @xmath165 and @xmath166 , the general term in the second summand is @xmath167 , where @xmath168 and @xmath169 , and in order for two such terms to have matching powers , for even only @xmath123 , @xmath170 should be the same as @xmath171 which would imply @xmath172 contradicting @xmath168 . therefore , to get the number of basic classes , we simply add the number of terms in each of these summands .
[ cor : non - isotopy ] there is no self - diffeomorphism of @xmath0 that maps one element of the family of tori @xmath173 to another . in particular , these tori are pairwise non - isotopic .
since the seiberg - witten invariant is a diffeomorphism invariant , so are the number of basic classes . on the other hand , by corollary [ cor : basics ] ,
the number of basic classes of @xmath109 is @xmath174 and for fixed @xmath175 this polynomial of @xmath176 is obviously increasing when @xmath177 . hence @xmath109 is diffeomorphic to @xmath178 if and only if @xmath179 proving that the tori in @xmath173 are different up to isotopy and in fact even up to self - diffeomorphisms of @xmath0 . this concludes the proof of theorem [ theorem : main ] .
we can easily extend theorem [ theorem : main ] to @xmath180 @xmath181 and more generally to @xmath182 , where @xmath112 is a symplectic 4-manifold satisfying certain topological conditions as in @xcite .
[ theorem : generalization ] assume that @xmath183 is a symplectic @xmath3-torus in a symplectic @xmath7-manifold @xmath112 .
suppose that @xmath184\in h_2(x)$ ] is primitive , @xmath184\cdot[f]=0 $ ] , and that @xmath183 lies in a fishtail neighborhood .
if @xmath185 , then we also assume that @xmath186 and is a finite sum .
then there exists an infinite family of pairwise non - isotopic symplectic tori in @xmath182 representing the homology class @xmath187\in h_2(x_p)$ ] for any integer @xmath5 .
these tori are inequivalent under self - diffeomorphisms of @xmath182 .
morton : the multivariable alexander polynomial for a closed braid , in _ low - dimensional topology _ , ed .
hanna nencka , _ contemporary mathematics _ * 233 * , amer . math .
( 1999 ) , 167172 . also available at arxiv : math.gt/9803138 .
sikorav : the gluing construction for normally generic @xmath191-holomorphic curves , in _ symplectic and contact topology : interactions and perspectives _
( toronto , on / montral , qc , 2001 ) , 175198 , fields inst .
commun . * 35 * , amer .
soc . , 2003 .
also available at arxiv : math.sg/0102004 . | let @xmath0 denote the rational elliptic surface with a single multiple fiber @xmath1 of multiplicity @xmath2 .
we construct an infinite family of homologous non - isotopic symplectic tori representing the primitive @xmath3-dimensional homology class @xmath4 $ ] in @xmath0 when @xmath5 . as a consequence
, we get infinitely many non - isotopic symplectic tori in the fiber class of the rational elliptic surface @xmath6 .
we also show how these tori can be non - isotopically embedded as homologous symplectic submanifolds in other symplectic @xmath7-manifolds . |
the coexistence of magnetism and superconductivity in heavy fermion compounds is one of the central issues in condensed matter physics@xcite . in particular , urhge and ucoge
have attracted much attention in recent years because they show the coexistence of long - range ferromagnetic order and superconductivity under an ambient pressure@xcite .
they are weak ferromagnets , with @xmath3 k and @xmath4 @xmath5 ( urhge ) , and @xmath6 k and @xmath7 @xmath5 ( ucoge ) .
they undergo transition into superconducting states below @xmath8 k ( urhge ) @xcite and @xmath9 k ( ucoge ) @xcite , and the superconducting state and magnetic orderings coexist below @xmath10 .
the natures of the superconductivity and magnetism in these compounds have been well studied experimentally .
for example , the magnetic field and the pressure phase diagrams of these compounds have been obtained experimentally , and it has been suggested that these compounds are located near the quantum critical point of the magnetic transition@xcite .
furthermore , the nqr study of ucoge clarified that the superconducting state and long - range magnetic ordering uniformly coexist in this compound @xcite . on the other hand , the electronic structures of these compounds are not well understood .
an itinerant nature of u 5@xmath1 states has been expected from the small ordering moments in the ferromagnetic phase as well as their pressure dependences @xcite .
in addition , x - ray photoelectron spectroscopy studies of urhge @xcite , urh@xmath11ru@xmath12ge @xcite , and ucoge @xcite showed that the u 5@xmath1 state forms peak structures near the fermi level ( @xmath13 ) , suggesting that the u 5@xmath1 electrons in these compounds have an itinerant nature .
however , precise information about u 5@xmath1 states has not been obtained in these studies , and a definitive conclusion has not yet been obtained . to establish realistic models of superconductivity and magnetism in these compounds , it is essential to reveal their microscopic electronic structures . here
we report the direct observation of the band structure of urhge through an angle - resolved photoelectron spectroscopy ( arpes ) study .
the band structure of urhge is obtained for the first time , and it is found that the u 5@xmath1 electrons form fermi surfaces ( fs s ) in this compound .
photoemission experiments were performed at the soft x - ray beamline bl23su of spring-8 @xcite .
the overall energy resolution in angle - integrated photoemission ( aipes ) experiments at @xmath14=800 ev was about 110 mev , and that in arpes experiments at @xmath15 ev was 100160 mev , depending on the photon energies .
the position of @xmath13 was carefully determined by measurements of the evaporated gold film .
clean sample surfaces were obtained by cleaving the sample _ in situ _ with the surface parallel to the @xmath16 plane .
the position of arpes cuts was calculated by assuming a free - electron final state with an inner potential of @xmath17=12 ev .
= 800 ev . ( a ) valence - band spectra of urhge and calculated partial density of states of the rh 4@xmath18 and u 5@xmath1 states .
( b ) u 4@xmath1 core - level spectrum of urhge together with those of the typical itinerant compound ub@xmath19 and localized compound upd@xmath20 . ]
first , we present the aipes spectra of urhge .
figure [ aipes ] ( a ) shows the valence - band spectrum of urhge taken at @xmath14=800 ev .
the sample temperature was kept at 20 k , and the compound is in a paramagnetic phase .
the spectrum is identical to that in ref .
@xcite . in this photon energy range ,
the contributions from u 5@xmath1 and rh 4@xmath18 states are dominant , and those from @xmath21 and @xmath22 states are two or three orders of magnitude smaller than those of u 5@xmath1 and rh 4@xmath18 states @xcite . in the valence - band spectrum , there is a sharp peak structure just below @xmath13 . on the high - binding - energy side
, there is a broad peak structure distributed at @xmath23 ev . to understand the origin of this peak structure ,
we have compared this spectrum with the result of the band - structure calculation . in the calculation ,
relativistic - linear - augmented - plane - wave ( rlapw ) band - structure calculations@xcite within the local density approximation ( lda)@xcite were performed for urhge treating all u 5@xmath1 electrons as being itinerant . in the lower part of the fig .
@xcite ( a ) , the calculated u 5@xmath1 and rh 4@xmath18 density of states broadened by the instrumental resolution are indicated .
comparison between the spectrum and the calculated density of states suggests that the sharp peak structure near @xmath13 and the broad peak structure on the high - binding - energy sides correspond to u 5@xmath1 and rh 4@xmath18 states , respectively . here
, it should be noted that a shoulder - like structure is recognized at around @xmath24 ev in the experimental spectrum .
a similar structure is present in the calculated u 5@xmath1 dos as well , and this suggests that this shoulder structure originates not from the electron correlation effect , but from the band structure .
the overall spectral shape is consistent with the results of the band - structure calculation .
figure [ aipes ] ( b ) shows the u 4@xmath1 core - level spectrum of urhge together with those of the typical itinerant 5@xmath1 compound ub@xmath19 and localized 5@xmath1 compound upd@xmath20 .
these data are taken from ref.@xcite .
they show a spin - orbit splitting corresponding to u 4@xmath25 and u 4@xmath26 , and both of them have asymmetric line shapes .
the core - level spectrum of urhge has a relatively simple spectral line shape .
it is similar to that of the itinerant 5@xmath1 compound ub@xmath19 , suggesting that u 5@xmath1 electrons in urhge have an itinerant character .
meanwhile , the spectrum of urhge is much broader than that of ub@xmath19 , and is accompanied by a satellite structure on the high - binding - energy side , as has been observed in the upd@xmath20 spectrum .
this implies that the u 5@xmath1 electrons in urhge have a correlated character as well .
we first explain about the brillouin zone and the calculated fss of urhge .
figure [ 3dfs ] shows the simple orthorhombic brillouin zone of urhge and the calculated fss in the paramagnetic phase . in the band - structure calculation ,
bands 6972 form fss as shown in the figure .
fs 69 is a small hole pocket around the y point .
fs 70 has a highly three - dimensional shape .
it forms a connected hole fs along the @xmath27 direction and a hole pocket fs in the middle of the @xmath28 and x points .
fs 71 is an electron fs with a grid - like shape spreading along the @xmath29@xmath30 plane .
fs 72 has a pillar - like shape along the @xmath30 direction at the corner of the brillouin zone .
all these calculated fss have large contributions from u 5@xmath1 states .
the calculation suggests that the electronic structure of this compound is highly three - dimensional in nature .
figure [ arpes1 ] shows the arpes spectra of urhge measured along the x-@xmath28-x [ fig .
[ arpes1](a ) ] , u - z - u [ fig .
[ arpes1](b ) ] , and s - y - s [ fig . [ arpes1](c ) ] high - symmetry lines .
the sample temperature was kept at 20 k , which was in the paramagnetic phase .
the position of the arpes cut in the momentum space is calculated based on the free electron final sates , and the photon energies used were @xmath31=625 ev for the x-@xmath28-x and u - z - u lines , and @xmath31=700 ev for the s - y - s line . in these arpes spectra , clear energy dispersions were observed . in the vicinity of @xmath13
, there exist dispersive bands with strong intensities .
these are contributions from the u 5@xmath1 quasiparticle bands .
the peak intensities of these structures have strong momentum dependences , suggesting that the u 5@xmath1 quasiparticle bands have finite energy dispersions . on the high - binding - energy side ( @xmath32 ev ) , there exist dispersive bands with strong intensities .
these are the contributions from the rh 4@xmath18 bands . here
, it should be noted that those bands are not symmetric relative to the high - symmetry points .
for example , the spectra measured along the x-@xmath28-x high - symmetry line are not symmetric relative to the x point in fig .
[ arpes1 ] ( a ) .
this may be due to the photoemission structure factor effect as has been observed in arpes spectra of other materials @xcite . to eliminate this effect , we have symmetrized these arpes spectra relative to the high - symmetry points .
figures [ arpes2 ] ( a)-[arpes2](c ) show the symmetrized arpes intensity maps .
the behaviors of the bands are clearly demonstrated in these images .
to evaluate the validity of the itinerant description of the u 5@xmath1 states in this compound , we compare the present arpes spectra with the result of the band - structure calculation treating all u 5@xmath1 electrons as being itinerant .
figures [ arpes2 ] ( d)-[arpes2 ] ( f ) show the calculated band structure . the color coding is the projection of the contributions from u 5@xmath1 states and rh 4@xmath18 states respectively . the contributions from u 5@xmath1 states are distributed in the energy range of @xmath33 ev , while those from rh 4@xmath18 states are mainly distributed in the energy range of @xmath34 ev .
many dispersive bands exist in the calculation , and it is difficult to compare them with the experimentally observed bands one by one . meanwhile , the overall band structures have some similarities between the experiment and the calculation . on the high - binding - energy side ( @xmath35 ev ) , the contributions from rh 4@xmath18 states were observed in both the experiment and the calculation . in the vicinity of @xmath13
, there are weakly dispersive bands in both the experiment and the calculation , and these are contributions from u 5@xmath1 states .
there exist few bands in the energy region in between them ( @xmath36 ev ) , and some calculated bands seem to correspond to the experimental spectra .
for example , there is an inverted parabolic band centered at the y point in the energy region of @xmath37 ev in the calculation [ fig .
[ arpes2 ] ( f ) ] that exists in the experiment [ fig .
[ arpes2 ] ( c ) ] .
the similar inverted parabolic bands centered at the @xmath28 point in the energy region of @xmath38 ev ( fig . [ arpes2 ] ( d ) ) are seen in the experimental spectra [ fig .
[ arpes2](a ) ] . to see details of the band structure near @xmath13 as well as its correspondence to the band - structure calculation , a blowup of the experimental arpes spectra and their simulation based on band - structure calculation
are shown in fig .
[ arpes3 ] .
figure [ arpes3 ] shows a comparison between experimental arpes spectra and their simulations along the x-@xmath28-x [ fig .
[ arpes3 ] ( a ) ] , u - z - u [ fig .
[ arpes3 ] ( b ) ] , and s - y - s high - symmetry lines [ fig .
[ arpes3 ] ( b ) ] .
these spectra are divided by the fermi - dirac function broadened by the instrumental energy resolution to observe the states near @xmath13 more clearly .
the approximate positions of experimental bands are estimated by the second derivatives of the energy distribution curves or momentum distribution curves , and are shown by dashed lines along the u - z - u high - symmetry line [ fig .
[ arpes3 ] ( b ) ] and the s - y - s high - symmetry lines [ fig .
[ arpes3 ] ( c ) ] . in the simulation ,
the following effects were taken into account : ( i ) the broadening along the @xmath30 direction due to the finite escape depth of photoelectrons , ( ii ) the lifetime broadening of the photo - hole , ( iii ) the photoemission cross sections of orbitals , and ( iv ) the energy resolution and the angular resolution of the electron analyzer .
the details are described in ref .
@xcite .
the correspondence between the arpes spectra and the calculations is more clearly recognized .
a detailed comparison suggests that there are some similarities between the experiment and the calculation .
some agreements are clearly identified , especially in the spectra along the u - z - u and s - y - s high - symmetry lines .
for example , along the u - z - u high - symmetry line , the experimentally observed three bands b , c , and d , correspond to the calculated bands 6566 , 6768 , and 6970 , respectively . in addition
, there are similar qualitative agreements in the spectra along the s - y - s high - symmetry line .
bands a , b , c , d , and e in the experimental spectra correspond to the bands 6364 , 6566 , 6768 , 69 - 70 , and 7172 , respectively .
in particular , band e forms a small electron pocket around the s point , which can be clearly recognized in the spectra normalized by the areas of the momentum distribution curves shown in the inset in the fig .
[ arpes3 ] ( c ) .
a similar electron pocket exists in the band - structure calculation as bands 71 and 72 . on the other hand
, band d does not have a large energy dispersion , and it does not form a fs as bands 69 and 70 do in the calculation . therefore , the agreement is unsatisfactory , especially in bands near @xmath13 .
the agreement is further unclear in the spectra along the x-@xmath28-x high - symmetry line .
the calculated spectra have a complicated structure , and it seems very different from the experimental spectra .
however , there are still some corresponding features in both the experiment and the calculation . for example , the inverted parabolic band with its apex at @xmath39 0.2 ev at the x point in the experimental spectra has a correspondence to bands 6364 in the calculation . the inverted parabolic band with its apex at @xmath39 1.0 ev at the @xmath28 point corresponds to part of the calculated bands 6165 .
the states near @xmath13 in the experimental spectra are rather featureless , and they are very different from the calculation . therefore , the agreement between the experiment and the calculation is limited , especially in states near @xmath13 , and we could not obtain information on the shape of fs s from the present experimental data .
in addition , many flat bands are expected around @xmath13 in the calculation , and even a tiny change in @xmath13 , by the order of 10 mev , drastically changes the shape of calculated fs s .
this makes it more difficult to compare states near @xmath13 between the experiment and the calculation .
next , we show the changes in the electronic structure due to the ferromagnetic transition .
figures [ tempdep ] ( a ) and [ tempdep ] ( b ) show blowups of arpes spectra measured along the x-@xmath28-x line at 20 k ( paramagnetic phase ) and 6 k ( ferromagnetic phase ) , respectively . although the changes are not significant , some temperature dependencies were clearly observed .
here , it should be noted that the energy difference between these two sample temperatures ( 20 k , @xmath401.7 mev ; 6 k , @xmath400.5 mev ) is much smaller than the energy scale of the changes in the arpes spectra , suggesting that the changes are not due to a thermal broadening effect . to see details of the changes
, we have subtracted the spectra measured at 20 k from those measured at 6 k. the intensity map of the difference in the spectra measured at 20 k vs 6 k is shown in fig . [ tempdep ] ( c ) , and the spectra at the @xmath28 point are depicted in fig .
[ tempdep ] ( d ) .
intensities are normalized with the highest intensity in the momentum and energy area shown in this figure .
both spectra are normalized to the intensities of the rh 4@xmath18 bands located at @xmath41 ev .
there are two kinds of changes in the spectral functions .
first , the intensities in the region of @xmath42 ev , which correspond to contributions mainly from u 5@xmath1 quasiparticle bands , decrease in the ferromagnetic phase .
the area is designated a in the figure .
the changes in spectral intensities near @xmath13 suggest that the ferromagnetic ordering in this compound originates from the itinerant u 5@xmath1 quasiparticle bands .
in addition to these changes , the intensities at around @xmath43 ev at the @xmath28 point decrease . the area is designated b in the figure .
the origin of this change is discussed in the next section .
accordingly , we have observed dispersive u 5@xmath1 quasiparticle bands near @xmath13 in arpes spectra of urhge .
they form fss in this compound , suggesting that they have basically an itinerant nature . on the other hand ,
the agreement between the experimentally obtained spectra and the band - structure calculation is limited .
the agreement is not satisfactory as those of the itinerant paramagnets ufega@xmath44 @xcite and ub@xmath19 @xcite , or the itinerant antiferromagnet un @xcite .
in particular , the states near @xmath13 ( @xmath2 ev ) show considerable deviations from the calculation .
the ellipsoidal pocket fs with a size of about 7 % of the brillouin zone was observed in sdh oscillations @xcite , but it does not exist in the calculated fs s .
this also suggests that the shape of fs s might be different from the calculation . these claim that the lda might be a reasonable starting point to describe its electronic structure , but inclusion of the electron correlation is needed to describe its electronic structure .
furthermore , the core - level spectrum of urhge has a satellite peak on the high - binding - energy side of the main line as shown in fig .
[ aipes ] ( b ) , suggesting the importance of the dynamical screening effect in urhge .
it has been shown that the inclusion of dynamical correlation effects alters the band structure near @xmath13 @xcite , and such a theoretical framework is a step forward to the understanding of this compound .
changes in the spectral line shape due to the ferromagnetic transition were observed in the vicinity of @xmath13 as well as on the high - binding - energy side ( @xmath45 ev ) .
the former change suggests that the ferromagnetic ordering in this compound is due to the changes in itinerant quasiparticle bands near @xmath13 .
although their details were not resolved in the present spectra , the changes are presumably due to the splitting of bands into majority - spin and minority - spin bands in the ferromagnetic phase as has been observed in ute @xcite and uir @xcite since u 5@xmath1 electrons have an itinerant nature in urhge as well . in the stoner - type mean - field model of itinerant ferromagnetism
, the exchange splitting energy @xmath46 is expressed as @xmath47 where @xmath48 and @xmath49 represent the stoner parameter and the magnetization respectively .
the stoner parameter @xmath48 might be somewhat different in the cases of ute and urhge , but the smaller magnetic moment of urhge ( @xmath4 @xmath5 ) compared to ute ( @xmath50 @xmath5 ) implies that the energy shift of the majority and minority bands ( @xmath51 ) of urhge should be much smaller than that of ute ( @xmath52 50 mev@xcite ) . furthermore , it is experimentally known that @xmath46 approximately scales with @xmath53 in itinerant ferromagnets@xcite , and the one order - lower transition temperature of urhge ( @xmath3 k ) compared to ute ( @xmath54 k ) also argues that its @xmath46 should be much smaller than that of uir .
these imply that the @xmath55 of urhge should be of the order of 10 mev , and its direct observation is difficult at the present energy resolution ( @xmath56 mev ) . meanwhile ,
the latter change on the high - binding - energy side is not the mean - field - like splitting of energy bands , but the change in peak structure .
this change might originate from the incoherent part of the spectrum due to the correlation effect of u 5@xmath1 states .
riseborough @xcite suggested that the incoherent spin excitation produces an incoherent peak in the off-@xmath13 region in a weak ferromagnet near a quantum critical point .
a similar change was observed in the spectra of the ferromagnet uir , where the peak at the @xmath57 point and @xmath580.5 ev was changed , in addition to the state near @xmath13 @xcite , by the ferromagnetic transition .
therefore , this change on the high - binding - energy side might be a common feature of uranium ferromagnets located near a quantum critical point .
in conclusion , we have found that quasiparticle bands with large contributions from u 5@xmath1 states form fss of urhge .
the overall band structure of urhge is explained by the band - structure calculation based on the lda , but the shape of the band structure near @xmath13 shows considerable deviations from the calculation .
the experimental band structure near @xmath13 is rather featureless , and the shapes of fs s are qualitatively different from the calculation . in addition , the u 4@xmath1 core - level spectrum of urhge is accompanied by a satellite peak , which is a signature of the electron correlation effect .
these results suggest that the inclusion of an electron correlation effect is essential to describe its electronic structure although the band - structure calculation is an appropriate starting point .
the changes in arpes spectra associated with the ferromagnetic transition were observed in the band near @xmath13 as well as states on the high - binding - energy side .
the former should be due to the exchange splitting of quasiparticle bands , while the latter might be related to the correlated nature of u 5@xmath1 states .
we would like to thank p. riseborough for stimulating discussion and comments .
the experiment was performed under proposal no .
2010b3824 at spring-8 bl23su .
the present work was financially supported by a grant - in - aid for scientific research from the ministry of education , culture , sports , science , and technology , japan , under contact no .
21740271 ; grants - in - aid for scientific research on innovative areas `` heavy electrons '' ( nos .
20102002 and 20102003 ) from the ministry of education , culture , sports , science , and technology , japan ; and the shorei kenkyu from hyogo science and technology association .
s. fujimori , t. ohkochi , i. kawasaki , a. yasui , y. takeda , t. okane , y. saitoh , a. fujimori , h. yamagami , y. haga , e. yamamoto , y. tokiwa , s. ikeda , t. sugai , h. ohkuni , n. kimura and y. nuki , j. phys .
* 81 * , 014703 ( 2012 ) .
a. yokoya , t. sekiguchi , y. saitoh , t. okane , t. nakatani , t. shimada , h. kobayashi , m. takao , y. teraoka , y. hayashi , s. sasaki , y. miyahara , t. harami and t.a .
sasaki , j. synchrotron rad .
* 5 * , 10 ( 1998 ) . t. durakiewicz , c. d. batista , j. d. thompson , c. g. olson , j.j .
joyce , g. h. lander , j. e. gubernatis , e. guziewicz , m. t. butterfield , a. j. arko , j. bona , k. matternberger , and o. vogt , phys .
lett . * 93 * , 267205 ( 2004 ) .
h. yamagami , t. ohkochi , s. fujimori , t toshimitsu , a. yasui , t. okane , y saitoh , a. fujimori , y. haga , e. yamamoto , s. ikeda and y. onuki , j. phys . : conference series * 200 * , 012229 ( 2010 ) . | the electronic structure of the ferromagnetic superconductor urhge in the paramagnetic phase has been studied by angle - resolved photoelectron spectroscopy using soft x rays ( @xmath0=595 - 700 ev ) .
dispersive bands with large contributions from u 5@xmath1 states were observed in the arpes spectra , and form fermi surfaces . the band structure in the paramagnetic phase
is partly explained by the band - structure calculation treating all u 5@xmath1 electrons as being itinerant , suggesting that an itinerant description of u 5@xmath1 states is a good starting point for this compound . on the other hand ,
there are qualitative disagreements especially in the band structure near the fermi level ( @xmath2 ev ) .
the experimentally observed bands are less dispersive than the calculation , and the shape of the fermi surface is different from the calculation .
the changes in spectral functions due to the ferromagnetic transition were observed in bands near the fermi level , suggesting that the ferromagnetism in this compound has an itinerant origin . |
in this paper we address the hot topic of the possible scale at which supersymmetry could manifest itself by performing indirect analyses based on the running of gauge couplings in two ways : * first a - by now `` traditional '' - analysis of gauge coupling unification in a grand unification theory , using the latest experimental inputs , mainly at the z peak scale , combined with a detailed statistical approach * a novel analysis , including for the first time high precision data collected at lep2 , which can shed light on the actual running of couplings from the z peak to @xmath3 207 @xmath1 . in the first approach in contrast to the sm the minimal supersymmetric standard model ( mssm ) leads to a single unification scale of a grand unified theory ( gut ) , if we let the couplings run accordingly .
the relevant parameters are : @xmath4 - a single generic susy scale where the spectrum of supersymmetric particles starts to play a role , and @xmath5 - the scale of grand unification where the electromagnetic , weak and strong coupling come together as an unified coupling @xmath6 .
we update the analysis @xcite , using the same fitting procedure and experimental inputs except for the stong coupling where we incorporate new data .
different measurements and world averages typically fall in two groups which we will call `` high '' and `` low '' values : + @xmath7 @xcite + @xmath8 @xcite + @xmath9 @xcite + @xmath10 @xcite . the results of the fits are summarized in table 1 .
@lclll@ inputs & threshold correction & @xmath4 & @xmath5 & @xmath11 + & [ % ] & [ @xmath1 ] & [ @xmath1 ] & + + s. bethke 2004 & - 4 & @xmath13 & @xmath14 & @xmath15 + @xmath16 & - 4 & @xmath17 & @xmath18 & @xmath19 + aleph 4-jet rate & - 4 & @xmath20 & @xmath14 & @xmath21 + + @xmath22 & - 4 & @xmath23 & @xmath24 & @xmath25 + the strong coupling is a key for interpreting the results : the `` high '' @xmath26 values require an uncomfortably low susy scale @xmath27 10 @xmath1 - way below the experimental lower limit @xmath27 100 @xmath1 .
the `` low '' @xmath26 values favor values for the susy scale just above the present limits , e.g. for @xmath26 = 0.1182 we get @xmath28 at one - sided 95 % cl - well in the lhc or even tevatron direct discovery range .
we perform a new more `` direct '' analysis by adding more data .
lep2 has measured the @xmath29 cross section above the z pole @xcite with combined precision from the four experiments @xmath27 1 % and theoretical uncertainty @xmath30 0.3 % .
we fit the data computing the running of the couplings from the z peak to the highest lep2 energies in the sm or in mssm with the scale as free parameter .
all three couplings change in a different way in the two cases if the susy scale is in this mass range .
as all couplings affect the theory prediction for @xmath31 , they could give measurable effects at this level of precision .
the slight `` excess '' @xmath27 2 standard deviations is `` filled '' nicely if the couplings run as predicted in mssm .
practically all high precision points are slightly above the sm prediction as could be the case if susy starts to take over - see figure [ fig : qqfit ] .
of course a more mundane explanation , which can not be ruled out at the current level of precision and understanding , is that we have a normalization problem either on the theoretical or experimental side .
the scale is fitted to be @xmath32 precision measurements of the strong coupling @xmath33 above the z pole can provide new information .
this coupling changes the fastest in this energy range , but is also the hardest to measure .
systematic uncertainties are dominant both at lep2 ( theory and hadronization ) and cdf ( energy scale and resolution ; interplay with the gluon parton density function uncertainty at high x values ) .
as shown in figure [ fig : alfas ] the experimental precision starts to approach the effects predicted by using our best fit value for the susy scale from @xmath31 .
if the tevatron experiments benefit fully from the increased statistics in run 2 to control the systematics better and probe higher energy scales , they could possibly provide new independent evidence for the scale of supersymmetry . | the apparent unification of gauge couplings around 10@xmath0 @xmath1 is one of the strong arguments in favor of supersymmetric extensions of the standard model ( sm ) . in this contribution
two new analyses of the gauge coupling running , the latter using in contrast to previous studies not data at the z peak but at lep2 energies , are presented .
the generic susy scale in the more precise novel approach is @xmath2 , easily within lhc , and possibly even within tevatron reach .
address = physics department , university of florida , p.o .
box 118440 , gainesville , fl 32611 , usa |
the study of massive stars ( @xmath11 8 m@xmath10 ) is one of the most prolific areas of research in present day astronomy , a fact which can be attributed to the not - so - well understood processes and a range of theories ( see * ? ? ?
* and references therein ) .
massive star formation , though a rare event in itself , can have a significant impact on its natal environment through processes such as strong stellar wind and supernovae explosions , which transfer large energy and momentum to the surrounding environment .
features such as bright - rimmed clouds and h ii regions ( owing to their large output of uv photons ) are hallmarks of massive stars in a region .
sh2 - 297 @xcite is an optically visible ( see figure [ fig_region ] ) , classical galactic h ii region ( @xmath12 @xmath4 07@xmath1305@xmath1413@xmath15 , @xmath16 @xmath4 -12@xmath1719@xmath600@xmath18 ) , ionized by a massive star , and bounded by a cold , dark cloud to the west .
it is associated with the reflection nebula and is a part of canis major ( cma ) ob1 association . in the literature , the distance estimates for from 1 to 1.4 kpc @xcite . in the present study , we have adopted the value of 1.1 kpc from @xcite . @xcite established the boundaries of as @xmath19 and @xmath20 .
a host of features populate the association - more than 30 nebulae @xcite , the association containing a group of stars ( including ) embedded in a prominent reflection nebula @xcite , and the three interconnected h ii regions of sh2 - 292 , sh2 - 296 , and ( see * ? ? ?
* and references therein ) .
@xcite , based on the h i morphology @xcite , emission nebulosity , and the distribution & properties of high - mass ob sources in the region , postulated that a supernova remnant ( snr ) could have induced star formation in .
they estimated the age of the supernova shell to be about half a million years .
the snr hypothesis was supported by the work of @xcite - who estimated a similar age for most of the stellar sources in using optical and infrared studies ; and that of @xcite , whose linear polarization survey of is consistent with the model of supernova induced compression of an initially uniform magnetic field , though , it must be noted that stellar winds too can produce a shell like structure .
however , an alternate star formation hypothesis by @xcite - based on the study of gas velocities and the fact that uv fluxes from two hot stars ( hd 54662 and hd 53975 ) can account for the physical parameters of the shell - suggests that strong stellar winds or evolving h ii regions are the cause of star formation in the region .
@xcite and @xcite have suggested the same .
@xcite have investigated this region using two micron all sky survey ( 2mass ) data and concluded that it presents similar structural properties as typical young and low mass open clusters .
@xcite found a deeply embedded , very young source of intermediate mass , named as uyso1 ( unidentified young stellar object 1 ) , that is driving a high - velocity bipolar co outflow at the interface of the h ii region and the dark cloud .
it was further resolved into two protostars , uyso1a and uyso1b @xcite .
a recent herschel pacs and spire mapping of the region around these sources has identified five cool and compact far infrared ( fir ) sources , of masses of the order of a few m@xmath10 , hidden in the neighbouring dark cloud @xcite .
continuing with our multiwavelength investigations of massive star - forming regions @xcite , in this paper , we explore the local star formation process around region .
the presence of multiple facets ( like the ionizing star s radiation field , the nebulosity present in the region , and the dark cloud to the west ) necessitates a multiwavelength study to decipher the star formation scenario .
hence , we carried out a detailed study of this region in optical , deep nir , mid infrared ( mir ) , and radio wavelengths to learn about the central ionizing source and its effect on the region , the physical characteristics of the region , the spatial distribution of the ysos and the nature of some selected individual ysos . based on the study of stellar population , their evolutionary status , and the morphology of the region ,
we have tried to deduce possible star formation activity . in section [ section_obs_and_data_reduction ] ,
we describe our observations and the data reduction procedures . in section [ section_other_data_sets ]
, we have outlined the other data sets which were used in our study .
we discuss the morphology of the region from radio , infrared ( ir ) and submillimetre data in section [ section_morphology ] , and the results pertaining to the stellar sources in section [ section_stellar_population ] . in section [ section_starformation_scenario ]
, we discuss the star formation scenario in region .
we present our main conclusions in section [ section_conclusions ] .
image ( red ) , 6563 h@xmath7 image ( green ) and 5007 [ oiii ] image ( blue ) , observed using hct ( see section [ subsection_optical_photometry ] ) .
the black square marks the position of the ionizing star hd 53623 . ]
optical photometric observations were performed using the 2 m himalayan _ chandra _ telescope ( hct ) , hanle ( india ) , operated by the indian institute of astrophysics , bangalore .
the telescope has ritchey - chretien optics with an f/9 cassegrain focus .
both broadband and narrowband observations were performed using himalayan faint object spectrograph and camera ( hfosc ) mounted on the telescope .
hfosc has a site ( scientific imaging technologies inc . )
2k @xmath1 4k pixels ccd , with the central region of 2k @xmath1 2k being used for imaging . with a pixel scale of 0.3@xmath18
, this translates to a field of view ( fov ) of @xmath4 10@xmath6 x 10@xmath6 .
long and short exposure observations were taken in bessell _ v _ ( 600 s & 20 s ) and _ i _ ( 300 s & 10 s ) filters ( centered at @xmath12 @xmath21 07@xmath1305@xmath1413@xmath15 , @xmath16 @xmath21 -12@xmath1719@xmath603@xmath18 ) on 2009 october 21 . [ sii ] ( 6724 ) , h@xmath7 ( 6563 ) , & [ oiii ] ( 5007 ) filter observations ( centered at @xmath12 @xmath21 07@xmath1305@xmath1417@xmath15 , @xmath16 @xmath21 -12@xmath1719@xmath618@xmath18 ) , with an integration time of 300 s each , were done on 2010 january 18 .
along with the object frames , several bias frames , twilight flat - field frames , and the standard field sa98 @xcite were also observed .
the average seeing size ranged from @xmath4 2@xmath18 to 3@xmath18 full width half maxima ( fwhm ) .
the standard field was used to calculate the extinction coefficients so as to calibrate our ccd systems . the astrometric calibration , with a position accuracy better than 0.3@xmath18 ,
was done using the 2mass sources in the field .
for all the filters , frames of equal exposure time were co - added to improve the signal - to - noise ratio .
photometry was carried out only for _ v _ and _ i _ filters .
image reduction and analysis facility ( iraf ) data reduction package was used for the initial ccd image processing .
thereafter , using daophot ii package of midas , the point spread function ( psf ) photometry and the photometric measurements were performed .
a variable psf was made using several uncontaminated stars in the field .
finally , the magnitudes ( _ v _ and _ i _ ) were calibrated in the manner outlined by @xcite , with a precision better than 0.05 mag . in the present work ,
we have used only those sources which have magnitude errors @xmath22 0.15 mag ( @xmath4 3@xmath23 ) .
the optical spectroscopic observations were carried out using hct and the 2 m iucaa girawali observatory ( igo ) telescope at girawali near pune ( india ) , operated by inter university centre for astronomy & astrophysics ( iucaa ) , pune . the hct data was obtained on 2009 december 21 using hfosc , with the help of grism 7 ( 3500 to 7000 @xmath24 ) - which has a resolving power of 1200 and a spectral dispersion of 1.45 @xmath24 pixel@xmath25 . the spectrum of the star ( @xmath12 @xmath26 07@[email protected]@xmath15 , @xmath16 @xmath26 -12@[email protected]@xmath18 ; v=7.99 mag ) with an exposure time of 480 s , in addition to fear lamp arc spectrum ( for wavelength calibration ) and multiple bias frames , was obtained .
the spectrophotometric standard star g91b2b @xcite was also observed with an exposure time of 1200 s. spectral reduction was carried out using `` apall '' task in iraf reduction package .
finally , the flux - calibrated spectrum was obtained ( see figure [ fig_hct_spectrum ] ) .
to cross - verify the weak diagnostic lines in the object spectrum from hct , we also obtained a deeper spectrum ( for 1200 s ) for using igo telescope , on 2009 november 26 .
the telescope has a cassegrain focus with a focal ratio of f/10 .
iucaa faint object spectrograph and camera ( ifosc ) , with a 2k @xmath1 2k pixel ccd array , was used for the purpose .
grism 7 , with a range of 3800 - 6840 , resolving power of 1300 , and a spectral dispersion of 1.49 pixel@xmath25 was used .
multiple bias frames , halogen flat - field frames , and hene lamp arc spectrum ( for wavelength calibration ) were obtained along with the object spectrum .
the reduction was done using `` doslit '' task in iraf .
no flux calibration was done for this case .
finally , the normalized spectrum was obtained ( see figure [ fig_igo_spectrum ] ) . to identify the h@xmath7 emission line stars ,
two grism slitless spectra ( 420 s each ) - using grism 5 ( 5200 - 10300 ) and h@xmath7 broadband filter ( @xmath27 6563 , @xmath28 500 ) - were also obtained with hct on 2007 november 16 ( centered at @xmath12 @xmath21 07@xmath1305@xmath1423@xmath15 , @xmath16 @xmath21 -12@xmath1719@xmath600@xmath18 ) . as a reference , for the identification of stars
whose slitless spectra were obtained , two h@xmath7 broadband filter images of the field were taken as well . here , the first step was to average the two photometric h@xmath7 filter images and the two slitless spectra of the region to improve their signal - to - noise ratio .
the world coordinate system ( wcs ) coordinates were then implemented on the averaged photometric image by identifying a bright set of stars in the fov , from the usno catalogue , and finally synthesizing their j2000 coordinates and pixel coordinates from the image , using the generic iraf commands `` ccmap '' and `` ccsetwcs '' .
the averaged spectrum was then analyzed in ds9 image display device , and the stars with an h@xmath7 emission line ( detectable as an enhancement over the continuum ) were identified manually by careful visual inspection , with the help of photometric image of the field .
19 h@xmath7 emission line stars were identified .
the detection limit of this identification is @xmath4 3.3 in terms of equivalent width , with the faintest star having a _ v _ magnitude of 20.07 .
fifteen were located in our nir fov , and hence the nir magnitudes for the remaining 4 were taken from 2mass catalogue for further analysis .
the region of interest , being a star forming region , has considerable nebulosity and dark cloud to the west , and hence is obscured at optical wavelengths .
moreover , we need to identify and classify the embedded ysos which emit at longer wavelengths , hence we use nir observations .
nir observations , centered on @xmath12 @xmath21 07@xmath1305@xmath1411@xmath15 , @xmath16 @xmath21 -12@xmath1718@xmath651@xmath18 for the object , in @xmath29 ( @xmath30 m ) , @xmath31 ( @xmath32 m ) and @xmath33 ( @xmath34 m ) bands , were carried out on 2007 february 27 using the f/10 cassegrain 1.4 m infrared survey facility ( irsf ) telescope , south africa with sirius ( simultaneous infrared imager for unbiased survey ) - a three color simultaneous camera .
the camera is equipped with three 1k @xmath1 1k hgcdte arrays , with a fov of @xmath4 7.8@xmath6 @xmath1 7.8@xmath6 , and a pixel scale of 0.45@xmath18 .
further information about the instrument can be found in @xcite and @xcite .
for each band , 40 frames each , centered at 9 dithered positions , were taken .
the integration time for each frame was 10 s , giving a total integration time of 3600 s ( 9@xmath140@xmath110 ) in each band for the region .
the sky condition was photometric , with the average seeing size ranging from 1.35@xmath18 to 1.60@xmath18 fwhm during the observations .
the standard data reduction procedure - involving bad pixel masking , dark subtraction , flat field correction , sky subtraction , combination of dithered frames , adding wcs coordinates to the image and psf photometry - was carried out using iraf .
astrometric calibration was implemented using the 2mass sources in the field .
about 30 - 40 stars , with good profiles , were selected for calibration for each band , and an accuracy of better than @xmath35 0.04@xmath18 was achieved for our absolute position calibration . after the initial processing , photometry was carried out on a final fov of @xmath4 7.5@xmath5 7.5@xmath6 .
the psf photometry was done using `` allstar '' algorithm of `` daophot '' package in iraf .
around 11 to 13 isolated bright stars were taken to make a psf for each band .
subsequently , the final magnitudes were calibrated using the 2mass stars in the field , with an rms of @xmath4 0.07 mag for all the filters . to ensure good photometric accuracy , we restricted our catalogue to sources having uncertainty @xmath22 0.15 mag ( @xmath4 2@xmath23 ) , independent of brightness , in all the three bands .
also , it was found that the sources with @xmath33 magnitude @xmath36 8 were saturated in our catalogue ; hence their magnitudes , in all three filters , were replaced by the corresponding 2mass values .
we evaluated the completeness limits for our images by doing artificial star experiments .
stars of varying magnitudes were added to the images , and then it was determined what fraction of stars were recovered in each magnitude bin .
the recovery rate was greater than 90@xmath37 for magnitudes brighter than 17 , 17 , and 16 in @xmath29 , @xmath31 , and @xmath33 bands , respectively .
the recovery percentage fell to 85@xmath37 for @xmath33 magnitudes in 16 to 16.5 mag bin .
the change in completeness limit is negligible on using the photometric error cutoff at 0.15 mag .
the final calibrated catalogue was analyzed using the cc and cm diagrams ( see section [ section_stellar_population ] ) .
for this purpose , the required magnitudes and colors were transformed into cit ( california institute of technology ) system using the color transformation equations given on the caltech websitejmc/2mass / v3/transformations/ ] .
in addition to the target region , a control field centered at @xmath12 @xmath21 07@xmath1306@xmath1420@xmath15 , @xmath16 @xmath21 -12@xmath1701@xmath640@xmath18 ( about 24@xmath6 to the north - east of our object field ) was also observed .
it was reduced in a similar manner as the object frames .
the control field was used to assess the level of contamination from foreground sources in the target field .
it was used to delineate the maximum @xmath38 color for the foreground sources , so as to identify the young stellar sources in the region - which will have a larger color by virtue of extinction and the presence of circumstellar disk which results in excess _ k_-band emission .
figure [ fig_nir_colorcomposite ] shows a false nir color composite ( _ j _ : blue , _ h _ : green , @xmath33 : red ) image of sh2 - 297 region , which reveals nebulosity around the central and south - eastern region , and the dark cloud towards the west .
( red ) images .
white square marks the position of ionizing star hd 53623 , white cross shows the position of iras 07029 - 1215 , and the magenta plus marks the location of uyso1 from @xcite . ]
1280 mhz observation was carried out on 2007 july 12 , while that for 610 mhz on 2007 july 19 , using giant metrewave radio telescope ( gmrt ) array .
the gmrt array consists of 30 antennae , arranged in an approximately y - shaped configuration , with each dish being 45 m in diameter .
12 antennae are located randomly in a compact 1 km@xmath11 km central square area , while the remaining 18 are located along the three radial arms ( 6 along each arm ) , each arm length being @xmath4 14 km .
the maximum baseline is @xmath4 25 km .
further details can be found in @xcite . for our observations , vla phase calibrator `` 0735 - 175
'' and flux calibrator `` 3c147 '' were used , at both the frequencies .
the details of the observations are given in table [ table_radio_data ] .
c c c date of obs . &
2007 july 12 & 2007 july 19 + phase center & @xmath12 = 07@[email protected]@xmath41 & @xmath12 = 07@[email protected]@xmath41 + & @xmath16 = -12@[email protected]@xmath18 & @xmath16 = -12@[email protected]@xmath18 + flux calibrator & 3c147 & 3c147 + phase calibrator & 0735 - 175 & 0735 - 175 + cont . bandwidth & 16 mhz & 16 mhz + primary beam & 24@xmath6 & 43@xmath6 + resolution of maps & & + used for fitting & [email protected]@xmath18 & [email protected]@xmath18 + peak flux density & 19.42 mjy / beam & 41.33 mjy / beam + rms noise & 2.32 mjy / beam & 3.24 mjy / beam + integrated flux density & 0.67 jy & 0.64 jy + [ table_radio_data ] the radio data from gmrt was analyzed with the help of astronomical image processing software ( aips ) distributed by nrao .
the data was edited to flag the bad antenna pairs , baselines , and time ranges using the `` vplot '' , `` uvflg '' , and `` tvflg '' tasks .
the final image was produced using a set of tasks which involved fourier inversion , cleaning ( using the task `` imagr '' in aips ) , followed by a few iterations of phase - only self - calibration ( till the rms convergence ) to remove the effects of the atmospheric and ionospheric phase distortions .
also , since our source is located near the galactic plane , we had to take into account the system temperature corrections for gmrt . since , at metre wavelengths , there is a large amount of radiation from the galactic plane , the effective temperature of the antennae increases , which has to be corrected for .
this was done by obtaining the sky temperature map of @xcite at 408 mhz , and rescaling the image by a scaling factor which is just the ratio of system temperature towards the target source and the flux calibrator ( here 3c147 ) .
the temperature for the source will be t@xmath44 + t@xmath45 , where t@xmath44 = t@xmath46 @xmath1(f/408)@xmath47 ( using the spectral index given in * ? ? ? * ) , with t@xmath46 being the temperature value from haslam s map at 408 mhz and `` f '' being the frequency ( in mhz ) for which the correction is to be calculated .
t@xmath45 , the system temperature , was obtained from the national centre for radio astrophysics ( ncra ) websitegtac / gmrt - specs.pdf ] for both the frequencies .
the temperature for the flux calibrator will merely be t@xmath45 , hence the scaling factor is ( t@xmath44 + t@xmath45)/t@xmath45 .
in addition , for 1420 mhz , the archival image from very large array ( vla ) , for the observation date 1998 december 14 , was also used ( project i d af346 ) .
radio continuum interferometric data , at 610 mhz , 1280 mhz and 1420 mhz was used to probe the ionized gas component .
the gmrt maps were convolved to a resolution of about 24@xmath4814@xmath18 , approximately same as the vla map . after obtaining the contour maps for all the three frequencies ,
the integrated flux was found by fitting a 3@xmath23 contour level box and using the aips task `` imstat '' ( see table [ table_radio_data ] ) .
using these radio continuum data points , the various physical parameters of the region were calculated ( see section [ subsection_physical_properties ] ) .
we used the 2mass data in the field for our photometric and astrometric calibrations .
the stars were obtained using the task `` imtmc '' in `` wcstools '' package in iraf , for the @xmath29 ( @xmath30 m ) , @xmath31 ( @xmath32 m ) and @xmath33 ( @xmath34 m ) bands .
all the sources which were obtained for calibration had a good quality flag value of `` aaa '' for the magnitudes .
we also substituted the 2mass magnitudes for those sources which were found to be saturated in our irsf catalogue , i.e. those for which we had irsf @xmath33 magnitude @xmath36 8 .
wide - field infrared survey explorer ( wise ) surveyed the entire sky in the nir bands at 3.4 @xmath49 m ( w1 ) and 4.6 @xmath49 m ( w2 ) , and in the mir bands at 12 @xmath49 m ( w3 ) and 22 @xmath49 m ( w4 ) , with an astrometric precision better than @xmath4 0.15@xmath18 @xcite .
the spatial resolution was @xmath4 6@xmath18 for the first three bands and @xmath4 12@xmath18 for 22 @xmath49 m .
since the wise ( preliminary release april 2011 ) catalogue had many sources missing ( e.g. for w1 band , only @xmath4 35@xmath37 of the sources detected in our photometry were present in the release catalogue ) , as could also be visually discerned , we carried out aperture photometry on the archival wise images - for the same fov as our irsf nir field - following the method described by @xcite .
this significantly improved the number of sources detected .
photometry was done using the daophot package in iraf ; with the aperture radius , inner radius of the sky annulus , & outer radius of the sky annulus being 5@xmath18 , 7.5@xmath18 , & 25@xmath18 for the 3.4 & 4.6 @xmath49 m bands , and 7.5@xmath18 , 10@xmath18 , & 25@xmath18 for the 12 @xmath49 m band , respectively . the 22 @xmath49 m band was not used mainly because of its low resolution , as well as the dominant nebulosity and the presence of not many point sources ( 12 sources were detected out of which 9 were in highly nebulous central region ) .
the photometry for each band was calibrated with the wise preliminary release catalogue .
care was taken not to use the objects flagged as `` d '' ( diffraction spikes ) , `` h '' ( halos from bright sources ) , `` p '' ( persistent artifacts ) , or `` o '' ( optical ghosts of bright sources ) in the release catalogue for calibration .
for the final wise catalogue , we took only those sources which had all magnitude errors @xmath50 0.3 , to make sure that they had reliable photometry .
our irsf nir catalogue was subsequently matched with the wise catalogues within a 3@xmath18 matching radius .
187 , 124 , and 9 sources detected in our nir fov had 3.4 @xmath49 m , 4.6 @xmath49 m , and 12 @xmath49 m counterparts , respectively .
the resultant final catalogue was used to get the seds of various sources , and hence in constraining the manifold physical parameters like age , mass , accretion rates etc . for them
( see section [ section_sed ] ) .
the 3.4 @xmath49 m and 12 @xmath49 m bands contain prominent polycyclic aromatic hydrocarbon ( pah ) features at 3.3 , 11.3 & 12.7 @xmath49 m @xcite in addition to the continuum emission , and hence can be used to get an idea of the photodissociation region .
the 22 @xmath49 m band can be used to examine the warm dust emission - the stochastic emission from small grains as well as the thermal emission from large grains @xcite .
the ionizing source of region is the star , associated with the reflection nebula @xcite . @xcite and
@xcite estimated the spectral type of this source as b1vn and b1ii / iii , respectively , using objective - prism plates from michigan curtis - schmidt telescope , ctio . @xcite
have proposed the spectral type as b0.5 iv - v using spectrogram analysis .
here we use the optical spectroscopic data to find out the spectral type of this source .
the reduced spectra from hct and igo , for the wavelength range 3950 to 4750 @xmath24 , have been shown in figure [ fig_spectra ] .
the spectra were analyzed with the help of walborn spectra catalogue @xcite .
the absence of he ii lines at 4200 & 4541 @xmath24 , and low strength at 4686 @xmath24 suggest that the star does not belong to o - type spectral class .
the most prominent lines in our spectra are the he i lines at 4009 , 4026 , 4144 , 4387 , 4471 , and 4713 @xmath24 .
the balmer lines at 3970 @xmath24 ( h@xmath51 ) , 4102 @xmath24 ( h@xmath52 ) , and 4341 @xmath24 ( h@xmath53 ) are clearly seen too .
we also see c iii + o ii blend at 4070 and 4650 @xmath24 .
weak si iv line at 4116 @xmath24 , weak o ii lines at 4415 - 4417 @xmath24 , and absence of si iii lines indicate that the star most probably does not belong to giant luminosity classes .
the preponderance of neutral helium lines and the comparison with the standard walborn catalogue suggest a most likely spectral type of b0v / b0.5v for the star . @xcite
have suggested , using h@xmath7 emission study , that this source does not contain any circumstellar disk , most probably due to photoevaporation by uv photons or due to influence of some supernova in the nearby region .
high resolution studies are important as they help us in understanding the structure of a region . has been observed with very coarse resolutions in previous radio surveys , which makes it harder to recognize the subtler features in the radio emission .
for example , the survey by @xcite observed 168 optically identified h ii regions at 1400 mhz , including , with a spatial resolution of 10@xmath6 .
similar observations were carried out by @xcite for wavelengths ranging from 2 to 11 cm , with resolutions of @xmath4 [email protected]@xmath6 . at larger frequencies ( 10 ghz ) ,
the cma r1 region was observed by @xcite with a resolution of 2.7@xmath6 .
@xcite , in a vla survey , observed the region at 1.4 ghz with 43@xmath18 resolution .
figure [ fig_radio_highresolution ] shows gmrt high resolution greyscale contour maps , for 610 mhz ( @xmath4 10@xmath48 5@xmath18 ) and 1280 mhz ( @xmath4 7@xmath48 6@xmath18 ) , of h ii region .
the intricate morphology of - showing the complex structures and the clumpy nature - can very well be discerned on examination .
the radio emission morphology from our observation as well as the location of h ii region at the tip of the molecular cloud ( see section [ subsection_ir_and_submm_structures ] ) suggest that the ionized gas is possibly undergoing a champagne flow ( also known as blister region ) @xcite , with the extended emission towards the east and the head of the ionized champagne flow located to the west towards ( see section [ section_starformation_scenario ] ) .
the radio continuum observations of the region using gmrt , and archival data from vla , were used to calculate the various physical parameters which characterize an h ii region ( emission measure , electron density , strmgren radius , and dynamical timescale ) . assuming a spherically symmetric and homogeneous ionized region , the model of @xcite for free - free emission was used . according to this , the flux density arising due to the region
is given by the following ( adapted from * ? ? ? * equations 1 and 3 ) : @xmath55 @xmath56 where , @xmath57 is the integrated flux density in jansky ( jy ) , @xmath58 is the electron temperature of the ionized core in k , @xmath59 is the frequency in mhz , @xmath60 is the electron density in @xmath0 , @xmath61 is the extent of the ionized region in pc , @xmath62 is the optical depth , @xmath63 is the correction factor , and @xmath64 is the solid angle subtended by the beam in steradian ( approximating the beam with a gaussian , the beam solid angle is given by @xmath65 , where @xmath66 and @xmath67 are major and minor half power beam widths ) .
the factor @xmath68 denotes the emission measure ( in @xmath3 pc ) , a measure of optical depth of the medium . for our calculations
, we set the value of @xmath63 as 0.99 ( obtained from * ? ? ?
* table 6 ) , and that of @xmath58 to 10000 k - applicable to classical h ii regions @xcite .
thereafter , treating the emission measure as a free parameter and using the three data points ( flux density vs. frequency ) from the three maps , a non - linear regression was implemented , following a procedure similar to @xcite ( see figure [ fig_radiofit ] ) , which yielded the value of the free parameter ( emission measure ) , for the best fit , as 9.15 @xmath35 0.56 @xmath1 10@xmath2 @xmath3 pc .
the extent of this extended h ii region is @xmath4 5@xmath6 , as obtained from the vla image . at a distance of 1.1 kpc
, this translates to an extent of 1.6 pc . from the emission measure and the extent ,
the electron density is obtained to be 756 @xmath35 46 @xmath0 .
the low values of the emission measure and the electron density reinforce the fact that this is a classical , and thus , evolved h ii region ( see * ? ? ?
* ) . 24@xmath4314@xmath18 .
] further , assuming a homogeneous and spherically symmetric nature for the h ii region , the luminosity of the lyman continuum photons ( in photons s@xmath25 ) from the central ionizing source is given by ( * ? ? ?
* equation 5 ) : @xmath69 where @xmath57 is the integrated flux density in mjy from the contour map , @xmath70 is the distance in kiloparsec , @xmath58 is the electron temperature , and @xmath59 is the frequency in ghz for which the luminosity is to be calculated . for our calculations , we used the 1420 mhz flux value ( 0.76 jy ) .
@xmath70 was taken to be 1.1 kpc and @xmath58 as 10@xmath71 k. the value for @xmath72 was determined to be 7.62@xmath110@xmath73 photons s@xmath25 ( hence @xmath74 is 46.88 ) .
@xcite have estimated the main sequence ( ms ) age of as 5 @xmath35 3 myr , therefore assuming a luminosity class v for it , a comparison of @xmath74 with ( * ? ? ?
* table ii ) yields the spectral type as b0 , in agreement with our optical spectroscopic analysis .
the lower value of our calculated luminosity ( @xmath74 ) as compared to the theoretical value ( 47.63 ) from @xcite could be because of absorption of photons by dust in the region .
as the central massive star ionizes the ambient medium around it , the ionization front moves outwards , and the whole h ii region expands till an equilibrium is reached between the number of ionizations and recombinations .
the underlying assumption here is that the ambient medium has homogeneous density and temperature . using the formulation given by @xcite , the radius of the h ii region at this point - called strmgren radius - is given by : @xmath75 where @xmath76 is the strmgren radius ( in cm ) , @xmath77 is the initial ambient density ( in @xmath0 ) , and @xmath78 is the total recombination coefficient to the first excited state of hydrogen .
the value of @xmath78 , for a @xmath58 of 10@xmath71 k , is taken to be 2.6@xmath110@xmath79 @xmath80 s@xmath25 @xcite .
we estimated the initial ambient density as follows ( similar to * ? ? ?
* ) . from the luminosity calculations for lyman continuum photons ,
we have obtained a calculated value of 7.62@xmath110@xmath73 photons s@xmath25 ( @xmath72 ) , which is about 18% of the theoretical output of 4.27@xmath110@xmath81 photons s@xmath25 @xcite .
this implies that @xmath82 , hence yielding @xmath77 as 4202 @xmath35 253 @xmath0 , using the above calculated value of the electron density . using these values for @xmath72 and @xmath77
, the strmgren radius turns out to be 0.051 @xmath35 0.002 pc .
as the gas expands , it sends a shock front into the neutral medium .
this shock front precedes the ionization front . at this stage ,
the radius of the h ii region is given by @xcite : @xmath83 where @xmath84 is the extent of the h ii region at time @xmath85 , and @xmath86 is the speed of sound , taken to be 11@xmath110@xmath2 cm s@xmath25 @xcite . using the current extent of 1.6 pc as the value of @xmath84 , and the above calculated value of @xmath76 in this equation , we obtain the value of @xmath85 , referred to as the dynamical timescale , as 1.07 @xmath35 0.31 myr .
it should be noted that , to calculate @xmath76 and @xmath85 , we have estimated @xmath77 ( 4202 @xmath35 253 @xmath0 ) assuming a luminosity class v for the massive star , and that only part of the radio flux is used for ionization .
also , we have used the current value of electron density . however , at the beginning of massive star formation , @xmath87 could be of the order of 10@xmath88 - 10@xmath89 @xmath0 @xcite . besides , the expansion velocity decreases with time . hence , the electron density and the dynamical timescale calculated here should be treated as conservative lower limits .
figure [ fig_co3to2 ] shows the @xmath90co(j=3 - 2 ) image of the region ( public processed data for observation date 2011 february 20 , project i d : m11ah48a ) from james clerk maxwell telescope ( jcmt ) . the msx 8.28 @xmath49 m contours ( in yellow ) , scuba 850 @xmath49 m contours ( in red ) @xcite , and the 1280 mhz radio contours from gmrt ( in white )
are overlaid .
the msx a band ( @xmath27 8.28 @xmath49 m ; containing pah features at 7.7 & 8.7 @xmath49 m besides continuum emission ) contours give an idea of the pah distribution in the region ( see e.g. * ? ? ?
@xmath90co(j=3 - 2 ) traces the molecular hydrogen , while scuba 850 @xmath49 m traces the cold dust .
as can be seen from the figure , there is a clear correlation between the ionized gas and 850 @xmath49 m contours in sh2 - 297 central region .
the @xmath90co image shows a d - shaped ring to the south of , which is also traced by the msx a band contours , and is faintly visible in narrowband h@xmath91 s(1 ) image of ( * ? ? ?
* see their figure 1 ) .
this is probably due to the sweeping up and piling of material from the expanding h ii region , and as such it is a potential site for future triggered star formation by collect - and - collapse process .
also noticeable is the fact that , towards the west , pah extends roughly upto the interface of the h ii region and ldn1657a .
co(j=3 - 2 ) image ( log greyscale with intensity range -1 to 30 k km s@xmath25 ) of the region from jcmt , overlaid with msx 8.28 @xmath49 m contours ( yellow ) showing the pah distribution , scuba 850 @xmath49 m contours ( red ) which trace the cold dust , and 1280 mhz radio continuum contours ( white ) indicating the ionized gas .
the co image reveals a maximum column density of 87.28 k km s@xmath25 .
msx and scuba contours have been plotted at 10 , 15 , 20 , 25 , 30 , 40 & 50 % of their peak values , which are 49.666 mw / m@xmath92-sr and 1.619 jy / beam , respectively .
radio contours ( @xmath4 25@xmath18 resolution ) have been plotted at 3 , 6 , 9 , 15 , and 21@xmath23 ( @xmath23 denotes the rms noise , which is 1.03 mjy / beam ) levels .
the blue square marks the ionizing star hd 53623 , blue plus symbol marks the position of uyso1 from @xcite , blue circle marks the position of c@xmath91h peak from @xcite , blue cross denotes iras 07029 - 1215 , and the sources from @xcite have been labelled and marked in white circles . ]
the wise w4 band which traces the warm dust and thermal emission from protostars ( not shown here ) , and scuba data show that the warm and cold dust extend well beyond the h ii region , and also towards the cold , dark cloud .
the gas and dust correlation is evident in the fir iras 100 @xmath49 m wavelength as well @xcite .
the nebula around the star shows a high level of fir emission .
@xcite found a c@xmath91h peak near , coinciding with uyso1 ( see figure [ fig_co3to2 ] ) , thereby indicating that this is a young star forming region @xcite .
@xcite have probed the dense molecular gas using the optically thin @xmath93 transition of @xmath94co .
one of the peaks is at @xmath95 , @xmath96 ( outside the range of figure [ fig_co3to2 ] ) , showing the presence of dense gas further towards the south - west of as well .
the nir cc diagram for the irsf sources , showing the @xmath97 vs. @xmath38 colors , is given in figure [ fig_nir_ccd_object ] .
the thick red line denotes the locus of ms stars , the magenta line shows the locus of giant stars while the dotted black straight line is the locus of classical t tauri stars ( ctts ) .
the loci for the ms and giants were taken from @xcite while those for the ctts from @xcite .
the three parallel dashed and slanted lines are the reddening vectors ( @xmath98 , @xmath99 and @xmath100 ) for the cit system from @xcite .
the crosses mark every 5 mag increase in @xmath101 .
we converted the colors from 2mass to cit system , as was discussed in section [ subsection_nir_observations ] .
we used the cc diagram to identify and study the nature of ysos in the region .
since young sources have a significant infrared excess , which in turn can be used to identify them , we divided our cc diagram into three regions - f , t and p ( similar to * ? ? ?
* ; * ? ? ?
the region `` f '' mostly contains the field stars and class iii sources ( weak - line t tauri stars ) .
the region `` t '' contains the ctts @xcite , class ii objects with large nir excesses , and reddened early type zams ( zero age main sequence ) stars with _ k_-band excess emission .
the `` p '' zone sources are most likely class i objects ( protostar - like objects ) with circumstellar envelopes .
the nir cc diagram is likely to be contaminated by the foreground and background field stars . to assess this
, we also plotted the cc diagram for the control field ( figure [ fig_nir_ccd_control ] ) . on comparison , we find that almost all our control field sources are confined to the left of the middle reddening vector , and a few which are to its right are below the ctt locus .
thus , the `` t '' and `` p '' regions of the object field are mostly uncontaminated by the field stars , and hence the sources in these regions can be assumed to be ysos .
as there is considerable nebulosity in the region , it must be kept in mind that there could be dust knots emitting at near and mid infrared wavelengths .
these could be mistakenly identified as class i / ii sources in the cc diagram . hence
, this cc diagram has its limitations when it comes to the identification of the ysos , and the information from it is only complementary to other techniques like spectroscopy , to understand the nature of objects .
the spatial distribution of the ysos identified was used to pin down the star formation process going on in this region ( see section [ subsection_spatial_distribution ] ) .
the nir cm diagrams , @xmath102 , of the region and the control field are shown in figure [ fig_nir_cmd ] .
the nearly vertical solid lines are the loci of zams reddened by @xmath103 and @xmath104 mag .
the slanting lines are the reddening vectors for the corresponding spectral types .
a comparison of the object field and the control field cm diagrams shows an apparent separation between the stellar sources at around @xmath105 0.6 .
our control field shows almost all sources having @xmath106 , thereby indicating that sources with @xmath106 are most likely the foreground stars .
the sources lying to the right of the @xmath107 cut - off line ( shown with a dashed vertical line in figure [ fig_nir_cmd_object ] ) have an excess emission in the _ k_-band and can thus reasonably be assumed to be young protostars . a cc diagram ( @xmath108/@xmath109 , * ? ? ?
* ) for the red sources which had wise counterparts ( not shown here ) showed their location in the diagram to be consistent with the location of class i sources .
the central ionizing star has been marked with a green square on this diagram .
the nir cm diagram indicates a b1 spectral type for the ionizing source , which is within one sub - class of that estimated from the optical spectroscopy and radio continuum observations .
however , a spectral type derived from just two data points ( @xmath31 and @xmath110 ) is not very reliable because of uncertainty in distance determination .
it just serves to give an idea of the spectral type , the spectrum analysis being canonical in this respect .
we estimated the extragalactic contamination ( extincted or faint background galaxies ) in our sample of nir excess sources .
we assume that the extragalactic sources can be seen through maximum extinction , which for our case is a@xmath111 25 mag ( a@xmath112 2.25 mag ) ( see section [ subsection_extinction ] ) .
accounting for the interstellar extinction and converting the _ _
k__@xmath113 magnitude to ab system ( since _ _ k__@xmath113 counts are tabulated in ab photometric system in @xcite ) , the total number of galaxy counts in our nir region , i.e. @xmath4 56.25 arcmin@xmath92 , turns out to be 6 @xmath35 2 down to our faintest detected magnitude of @xmath4 17 in the _ _ k__@xmath113 band .
a list of all the ysos , with their respective magnitudes in optical and ir , has been given in table [ table_ysos ] .
97 ysos were analyzed in total , with 93 of them in our nir fov of @xmath4 7.5@xmath5 7.5@xmath6 , and 4 h@xmath7 emission line stars outside this fov ( but considered for analysis nevertheless ) .
out of the 93 ysos , 15 are h@xmath7 emission line stars detected using grism slitless spectroscopy with a _
v _ limiting magnitude of 20.07 , 76 are nir ysos ( 26 class ii - type sources @xmath114 50 red sources with @xmath8 ) with a limiting @xmath33 magnitude of @xmath4 17 , and 2 are ldn1657a-2 and ldn1657a-3 sources from @xcite ( only ldn1657a-3 was detected in nir ) .
class ii - like and @xmath8 sources which were also detected as h@xmath7 emission line sources were classified as h@xmath7 emission line sources .
the @xmath8 sources which overlapped with class ii - like sources were classified as class ii - like sources .
c c c c c c c c c c c c c c c + the average extinction in the region was calculated using the cc diagram as well as the extinction map of the region . to calculate the average reddening using the cc diagram
, we used the sources in the `` f '' and `` t '' zones of our diagram .
the sources in the `` t '' region were dereddened to the ctt locus , while the sources in the `` f '' region were dereddened to the ms locus which was approximated by a straight line asymptote .
this gives the color - excess for each source .
subsequently , the reddening laws of @xcite were followed to calculate the @xmath101 for each source .
finally , we fit the set of @xmath101 values ( ranging from @xmath4 0.02 to 16 mag ) obtained for each of the sources , with a normal distribution .
the mean of this distribution was found to be 2.70 mag .
the average visual extinction for the region was also found by making an extinction map ( nice method , * ? ? ?
* ; * ? ? ?
to do this , we first obtained the intrinsic @xmath38 color from the control field .
thereafter , the color - excess , e(@xmath38 ) , and then @xmath101 for each of the sources in our catalog was calculated .
the color - excess is related to visual extinction by : e(@xmath38 ) = 0.065@xmath1 @xmath101 , which can be calculated from the reddening laws .
@xmath101 values ranged from 0@xmath925 mag , as we took all sources with at least _ h _ and _ k _ magnitudes , and thereby even redder ones ( undetected in _ j _ ) , here .
subsequently , the extinction map was made , where the extinction at each pixel was calculated by a weighted gaussian mean around it . for our case
, we took a gaussian with an fwhm of 1@xmath6 .
finally , the mean pixel value for our map was calculated using the task `` imstat '' in iraf , which gave us an average visual extinction of 2.69 mag .
hence , we find similar @xmath101 values from both the methods .
we used 2.70 mag as our value for @xmath101 in the optical cm diagram to estimate the mass and age of the ysos ( see section [ subsection_optical_cmd ] ) .
it gives the mean extinction of the field .
however , owing to the nebulosity in the central region and the presence of dark cloud to the west ( see section [ subsection_spatial_distribution ] ) , individual extinction for the sources will vary depending on their location .
figure [ fig_optical_cmd ] shows the _
v / v - i _ cm diagram for the ysos in the region . here
, h@xmath7 emission line sources have been shown in orange crosses , class ii - like sources in blue circles , and sources with @xmath8 have been represented by red circles .
most of the sources detected at optical wavelengths are h@xmath7 emission line sources , with only three class ii - like sources and one source with @xmath8 .
this is expected as h@xmath7 emission line sources are the most evolved ysos in the field . in figure
[ fig_optical_cmd ] , the isochrone for 2 myr for solar metallicity by @xcite , and pre - main sequence ( pms ) isochrones for 0.1 to 5 myr from @xcite have been overlaid .
the isochrones were corrected for distance ( 1.1 kpc ) and reddening ( @xmath101 = 2.70 mag ; see section [ subsection_extinction ] ) .
the figure also shows the evolutionary tracks for masses ranging from 0.3 to 3 m@xmath10 from @xcite .
emission line stars ( orange crosses ) , class ii - type objects ( blue circles ) , and @xmath8 source ( red circle ) . the thick solid line is the 2 myr isochrone by @xcite , while the thinner solid lines are pms isochrones for different ages by @xcite .
dashed lines are evolutionary tracks for different masses by @xcite .
the isochrones have been corrected for distance ( 1.1 kpc ) and reddening ( @xmath101 = 2.70 mag ) .
the arrow shows the reddening vector for @xmath101 = 3 mag . ]
we can see that most of the ysos lie in the age range of 0.5 to 2 myr , with the mean age @xmath4 1 myr .
the mass range mostly varies from 0.3 to @xmath4 2 m@xmath10 , with a few outliers mainly being the h@xmath7 emission line sources .
the pms star to the right of 0.1 myr isochrone is likely to be a highly extincted source .
it should be noted that variable extinction along the line of sight and variability could be the reason behind the age spread in our figure .
similar age spreads have been noted in other star - forming regions @xcite .
the excess _
k_-band emission can make the ysos appear luminous and hence the @xmath115 cm diagram is not appropriate for mass estimation .
we therefore choose the @xmath116 cm diagram to minimize the effect of this excess emission which would lead to an over estimation of mass .
figure [ fig_mass_spectrum ] shows the required cm diagram , with the class ii - like sources detected from the cc diagram ( in blue circles ) , sources with @xmath8 ( in red circles ) , and sources with h@xmath7 emission line ( in orange crosses ) .
the ionizing star has also been shown with a green square .
the larger statistics here as opposed to the optical cm diagram is due to the young nature of the sources , which emit at longer wavelengths . from the optical cm diagram
, we found that the average age of the ysos was about 1 myr .
also , the dynamical age of this h ii region from radio results is 1.07 @xmath35 0.31 myr ( see section [ subsection_physical_properties ] ) . hence , we use the isochrone for 1 myr ( black solid line ) from @xcite for our analysis .
the slanting dotted lines show the reddening vectors for different masses . .
the h@xmath7 emission line stars have been represented by orange crosses .
the central ionizing star has been marked by a green square .
the pms isochrone for 1 myr ( solid line ) from @xcite has been shown .
the isochrone was corrected for a distance of 1.1 kpc .
the slanted lines show the reddening vectors for different masses . ]
we can see from the 1 myr reddening vectors , that most of the h@xmath7 emission line objects and class ii - like sources lie between 0.1 to 2 m@xmath10 , though a few outliers seem to be much more massive , of the order of @xmath4 3 m@xmath10 .
most @xmath8 sources have masses less than 2 m@xmath10 .
it is noteworthy that our mass range is consistent with the mass range derived from the optical cm diagram ( section [ subsection_optical_cmd ] ) , with the h@xmath7 emission line stars being towards the massive end . as can clearly be seen , the ysos with the assumed age , have widely varying colors
this is probably an indication of variable extinction , weak contribution of excess emission in _
j _ and _ h _ bands , and sources being in different evolutionary stages .
it should be noted that mass estimations from infrared cm diagrams are prone to systematic errors @xcite due to possible use of different pms evolutionary models , and reliance on uncertain age and distance .
another error source could be the presence of binaries , which will brighten a star - thus changing the mass and age estimates .
so far , we have identified the ysos with the help of nir cc and cm diagrams , and the h@xmath7 emission line stars .
to get an idea about the physical parameters of these sources , we modelled the seds using the grid of models and fitting tools of @xcite .
the basic model assumes three components of any source : a pms central star , a surrounding flared accretion disk , and a rotationally flattened envelope with cavities carved out by a bipolar outflow .
different models are computed for a suitably large parameter space - containing manifold combinations of the above mentioned components - using a monte carlo based radiation transfer code @xcite .
sed fitting for different sources is not unique and the number of models for each source can only be constrained by having more data points from multiwavelength observations . to reduce error due to the uncertainties which result from very few data points ,
we modelled only those sources for which we had wise data and at least five data points .
the input parameters were distance range ( taken to be [email protected] pc ) , and a@xmath117 ( taken as 1.1@xmath925 mag ) .
the lower limit was chosen to be 1.1 mag ( from the calculation of @xcite for ) as the ysos should have an extinction which is more than the visual absorption for the region .
the upper limit was chosen as 25 mag because the highest individual a@xmath117 values from color - excess calculations using nice method ( section [ subsection_extinction ] ) were around this value .
the sed fitting tool of @xcite gives a set of well constrained values for each parameter , rather than the values for only the best fit for sed .
following a similar method as @xcite , we considered only those fits to constrain the physical parameters for which we had @xmath118 where @xmath119 is measure of goodness of fit for each model ( see figure [ fig_sed ] ) .
the final values were subsequently obtained using the models which satisfied equation [ equation_sed ] and finding a weighted mean value for each parameter , with the weight being the inverse square of @xmath119 for each model respectively .
for some of the sources , the errors for a few parameters are large because we are dealing with a large parameter space , while we have only a few data points to fit the models .
the results from our sed modelling are given in table [ table_sed ] .
the physical parameters listed are stellar mass ( @xmath120 ) , stellar age ( @xmath121 ) , temperature ( @xmath122 ) , disk accretion rate ( @xmath123 ) , foreground visual extinction ( @xmath101 ) , and the @xmath124 per data point of the best fit . c c c c c c c c c + 1 & 07:04:58.030 & -12:16:50.80 & 0.21 @xmath35 0.39 & 3.37 @xmath35 0.46 & 3.46 @xmath35 0.07 & -6.82 @xmath35 0.55 & 16.23 @xmath35 7.82 & 4.36 + 2 & 07:05:00.665 & -12:16:45.13 & 1.27 @xmath35 1.49 & 5.32 @xmath35 0.39 & 3.61 @xmath35 0.10 & -7.71 @xmath35 1.85 & 10.56 @xmath35 4.20 & 1.89 + + 19 & 07:05:00.919 & -12:15:42.37 & 2.44 @xmath35 1.00 & 5.43 @xmath35 0.44 & 3.67 @xmath35 0.05 & -9.15 @xmath35 1.97 & 7.27 @xmath35 1.14 & 1.40 + 45 & 07:05:10.395 & -12:19:25.70 & 2.91 @xmath35 1.38 & 4.55 @xmath35 1.27 & 3.81 @xmath35 0.27 & -6.35 @xmath35 1.45 & 1.89 @xmath35 0.78 & 8.28 + 49 & 07:05:13.678 & -12:19:51.96 & 1.14 @xmath35 0.72 & 5.05 @xmath35 0.38 & 3.61 @xmath35 0.03 & -8.18 @xmath35 1.11 & 2.90 @xmath35 1.83 & 1.20 + + 53 & 07:04:56.673 & -12:16:12.77 & 0.97 @xmath35 0.82 & 5.21 @xmath35 0.37 & 3.60 @xmath35 0.03 & -8.74 @xmath35 1.20 & 2.22 @xmath35 1.16 & 3.23 + 54 & 07:04:57.964 & -12:17:13.85 & 0.19 @xmath35 0.22 & 5.90 @xmath35 0.48 & 3.49 @xmath35 0.03 & -10.05 @xmath35 1.46 & 3.94 @xmath35 0.88 & 1.45 + 58 & 07:05:01.230 & -12:19:03.69 & 1.28 @xmath35 0.90 & 5.15 @xmath35 0.51 & 3.62 @xmath35 0.04 & -9.01 @xmath35 1.41 & 2.98 @xmath35 1.32 & 3.34 + 64 & 07:05:08.269 & -12:19:15.83 & 3.62 @xmath35 2.24 & 5.37 @xmath35 1.20 & 3.87 @xmath35 0.30 & -8.82 @xmath35 2.73 & 15.93 @xmath35 5.87 & 0.34 + 66 & 07:05:09.197 & -12:19:00.00 & 0.86 @xmath35 0.74 & 4.63 @xmath35 0.72 & 3.59 @xmath35 0.09 & -7.88 @xmath35 1.20 & 4.72 @xmath35 2.26 & 1.64 + 69 & 07:05:11.424 & -12:19:24.36 & 0.86 @xmath35 0.71 & 4.72 @xmath35 0.65 & 3.59 @xmath35 0.08 & -7.98 @xmath35 1.15 & 4.45 @xmath35 1.95 & 1.50 + 70 & 07:05:11.541 & -12:18:11.39 & 1.32 @xmath35 1.35 & 4.43 @xmath35 0.79 & 3.60 @xmath35 0.06 & -7.93 @xmath35 1.12 & 3.07 @xmath35 1.37 & 2.21 + 72 & 07:05:11.774 & -12:16:24.01 & 1.00 @xmath35 0.80 & 5.24 @xmath35 0.58 & 3.64 @xmath35 0.13 & -8.56 @xmath35 1.45 & 2.71 @xmath35 1.52 & 2.13 + 73 & 07:05:14.495 & -12:20:47.57 & 0.53 @xmath35 0.79 & 4.93 @xmath35 0.89 & 3.52 @xmath35 0.14 & -8.76 @xmath35 1.21 & 2.45 @xmath35 1.19 & 0.03 + 75 & 07:05:15.324 & -12:20:02.05 & 0.84 @xmath35 1.09 & 4.42 @xmath35 0.90 & 3.58 @xmath35 0.15 & -7.45 @xmath35 1.23 & 2.60 @xmath35 1.92 & 2.15 + 76 & 07:05:17.419 & -12:18:58.69 & 0.81 @xmath35 1.19 & 4.85 @xmath35 1.15 & 3.56 @xmath35 0.22 & -8.20 @xmath35 1.42 & 2.21 @xmath35 1.01 & 0.04 + + 79 & 07:05:06.330 & -12:15:36.38 & 2.27 @xmath35 0.67 & 6.63 @xmath35 0.36 & 3.85 @xmath35 0.15 & -9.95 @xmath35 1.95 & 5.11 @xmath35 1.05 & 1.91 + 80 & 07:05:09.652 & -12:19:56.32 & 0.89 @xmath35 1.18 & 4.60 @xmath35 0.82 & 3.58 @xmath35 0.06 & -7.90 @xmath35 0.98 & 3.24 @xmath35 1.59 & 1.15 + 82 & 07:05:12.292 & -12:18:38.27 & 0.94 @xmath35 0.92 & 5.68 @xmath35 0.30 & 3.61 @xmath35 0.04 & -8.76 @xmath35 1.15 & 1.46 @xmath35 0.52 & 1.52 + 86 & 07:05:17.904 & -12:15:40.26 & 4.60 @xmath35 1.03 & 5.61 @xmath35 0.07 & 3.78 @xmath35 0.09 & -7.48 @xmath35 1.00 & 2.01 @xmath35 0.60 & 2.62 + 87 & 07:05:18.494 & -12:15:09.22 & 2.30 @xmath35 0.72 & 6.04 @xmath35 0.58 & 3.79 @xmath35 0.17 & -9.08 @xmath35 1.49 & 2.31 @xmath35 1.05 & 0.54 + 88 & 07:05:19.389 & -12:20:28.71 & 2.51 @xmath35 0.81 & 6.55 @xmath35 0.31 & 3.82 @xmath35 0.16 & -9.41 @xmath35 2.01 & 2.23 @xmath35 0.92 & 0.24 + 89 & 07:05:19.592 & -12:17:31.48 & 2.33 @xmath35 0.86 & 5.58 @xmath35 0.42 & 3.67 @xmath35 0.05 & -9.61 @xmath35 1.32 & 2.12 @xmath35 0.58 & 2.68 + 90 & 07:05:20.068 & -12:19:12.59 & 3.93 @xmath35 0.15 & 6.63 @xmath35 0.13 & 4.15 @xmath35 0.01 & -12.90 @xmath35 0.60 & 3.83 @xmath35 0.21 & 1.27 + 91 & 07:05:20.728 & -12:21:42.69 & 2.64 @xmath35 0.78 & 5.57 @xmath35 0.31 & 3.68 @xmath35 0.05 & -9.15 @xmath35 1.64 & 1.91 @xmath35 0.66 & 1.04 + 92 & 07:05:21.912 & -12:19:07.19 & 0.26 @xmath35 0.18 & 5.83 @xmath35 0.29 & 3.51 @xmath35 0.02 & -10.70 @xmath35 1.64 & 1.34 @xmath35 0.34 & 2.61 + 93 & 07:05:22.359 & -12:17:18.63 & 1.07 @xmath35 0.88 & 6.14 @xmath35 0.45 & 3.62 @xmath35 0.06 & -9.20 @xmath35 1.18 & 2.06 @xmath35 1.00 & 1.25 + + 96 & 07:05:23.669 & -12:22:36.08 & 1.69 @xmath35 0.64 & 6.22 @xmath35 0.38 & 3.71 @xmath35 0.13 & -9.96 @xmath35 1.58 & 2.46 @xmath35 1.11 & 1.64 + two of the pms sources detected by @xcite - ldn1657a-2 , & ldn1657a-3 - had wise counterparts .
ldn1657a-3 was also detected at nir wavelengths . using the flux values given for these sources at longer wavelengths from @xcite
, we could further constrain the physical parameters of these sources .
the source with i d `` 1 '' in our sed table ( table [ table_sed ] ) gives the values for ldn1657a-2 , while `` 2 '' gives those for ldn1657a-3 .
figure [ fig_sed ] shows examples of seds with the resulting models for ldn1657a-3 and a typical class - ii like source .
it must be kept in mind that the sed results are only representative of the actual values and should not be taken too literally . out of a database of 200000 yso models ( each for a different set of parameter values ) ,
the model simply finds the closest matches for a particular source entered by the user .
any grid of models will have inherent limitations , which could give rise to pseudo - trends in the results .
the models are for isolated objects and thus could be misleading when we have unresolved multiple point sources .
moreover , the model assumes same physics for stellar masses from 0.1 to 50 m@xmath10 .
@xcite provides a succinct summary of the caveats inherent in this modelling .
nevertheless , we can still use the results to glean general information about the mass and age ranges of our sources . as can be seen from table [
table_sed ] , most of the sources have stellar ages in the range 0.05 to 1 myr and masses in 0.2 to 2.5 m@xmath10 range .
the outliers are expectedly the more evolved h@xmath7 emission line stars , whose ages vary upto 5 myr , and stellar masses are @xmath4 2.5@xmath95 m@xmath10 .
we see a few very low mass ( 3 sources with masses @xmath22 0.5 m@xmath10 ) and extremely young sources ( 9 sources with ages @xmath125 yr , out of which 6 are ctts ) as well .
the disk accretion rate for most of the h@xmath7 emission line sources is in the range of @xmath4 10@xmath126 - 10@xmath127 m@xmath10 yr@xmath25 , which is less than that of ctts @xmath4 10@xmath128 - 10@xmath126 m@xmath10 yr@xmath25 , though it would be prudent to treat these values with caution owing to a lack of fir and submillimetre data points .
a few of our sources , i d no.s 39 , 55 , 61 , 81 , 83 , 94 , 95 , and 97 ( table [ table_ysos ] ) , were better fitted by stellar models of a heavily extincted central star , and are thus not included here .
hence , we find that our age and mass estimations are mostly consistent with those derived from the optical cm diagram and the mass spectrum above .
figure [ fig_nir_colorcomposite_with_sources ] shows the spatial distribution of the ysos from our irsf and grism slitless spectroscopy data .
red circles denote the sources with @xmath129 , sources with @xmath130 have been shown with magenta crosses , blue circles denote the class ii - like objects detected from our cc diagram , while the white crosses show the location of the h@xmath7 emission line stars from grism slitless spectroscopy .
the central ionizing source is shown with a green square .
@xcite had found a young , intermediate mass protostar named uyso1 at the north - west border of the h ii region , towards the cold dark cloud .
the location of uyso1 , which was further resolved into two protostars @xcite , has been denoted by a green plus sign .
in addition , a recent herschel pacs and spire mapping of the region around uyso1 identified five cool and compact fir sources , of masses of the order of a few solar masses , hidden in the dark cloud @xcite .
they have been marked by green circles .
most of the ysos identified by us lie in the molecular cloud region traced by @xmath90co molecular emission ( see section [ subsection_ir_and_submm_structures ] & figure [ fig_co3to2 ] ) , along the concave arc facing the north - eastern corner .
( red ) band images .
the h@xmath7 emission line stars have been denoted by white crosses , class ii - like sources by blue circles , sources with @xmath130 with magenta crosses , and sources with @xmath129 by red circles .
the green square marks the position of the central ionizing star hd 53623 .
uyso1 @xcite has been marked with a green plus sign , while the cool , compact sources from @xcite by green circles .
1280 mhz gmrt contours ( @xmath4 25@xmath18 resolution ) showing the champagne flow have been shown in dashed white lines .
the three sub - regions denoted by the yellow boxes show an increasing fraction of redder sources in the sub - regions as we move from the ionizing star towards the north - western direction .
the dashed yellow lines are drawn at a distance of 80@xmath18 and 220@xmath18 from the ionizing star , respectively . ]
we have divided the region into three sub - regions which show increasing infrared excess from the south - east towards the north - west direction .
this indicates an evolutionary sequence , thereby suggesting a possible decreasing age gradient towards that direction .
the youngest stars , i.e. those with @xmath8 mostly lie towards the north - west direction , followed by a concentration of class ii - like objects and then the h@xmath7 emission line stars towards the ionizing region . even among the stars with @xmath8 , the redder sources ( @xmath129 )
were found to be lying towards the north - west , with a decreasing color sequence towards the ionizing source .
also , the 5 sources found by herschel mapping are cooler than the massive object uyso1 @xcite , thereby indicating that they are younger than uyso1 .
we tried to quantify the infrared excess ( @xmath38 ) for increasing distances of the sources from the ionizing star hd 53623 .
figure [ fig_distance_and_color ] shows a plot of @xmath38 color _ versus _ distance .
the graph is divided into three regions as shown by three boxes in figure 12 , with the average @xmath38 color calculated for each of them .
the average color was found to increase as we move away from the ionizing star , thereby indicating that the sources in general , are younger at larger distances from the central region .
color with distance from the star hd 53623 .
the h@xmath7 emission line stars have been denoted by black circles , class ii - like sources by blue circles , sources with @xmath130 with magenta crosses , and sources with @xmath129 by red circles .
green circle denotes ldn1657a-3 source . ]
this distribution of ysos , with relatively larger nir excess at larger radial distances , could be indicative of probable triggered star formation , which seems to have propagated from the ionizing source towards the north - west .
star formation in a molecular cloud is more active during the first few million years of cloud time .
subsequent to that , the gas in molecular cloud dissipates and further star formation no longer takes place .
however , these types of molecular cloud complexes can still have star formation at the interface of the h ii region created by ob stars and the molecular gas . in region ,
the spatial distribution of the ysos and infrared excess calculations ( section [ subsection_spatial_distribution ] ) suggest an evolutionary sequence of sources from the ionizing region towards the cold , dark cloud , with a majority of ysos distributed at the western and north - western border of .
one of the conditions for the massive star to have influenced the formation of ysos in the region is that the age of the h ii region and associated sources should be of the order of , or greater than , the age of the ysos in the triggered region .
however , it is difficult to estimate the age of ysos as very few of them have optical counterparts .
nonetheless , we tried to estimate the age of the few sources which have optical counterparts in section [ subsection_optical_cmd ] .
age estimation using ir data could be misleading because ir emission arises not only from the photosphere , but also the circumstellar disk of the ysos .
we tried to constrain the age of embedded ysos with the sed fitting models , though due to the lack of sufficient data points , the estimated ages are uncertain .
however , they can still be used as quantitative indicators of stellar youth .
our analysis shows the mean age of ysos to be @xmath4 1 myr , which is of the order of the dynamical age of the h ii region of 1.07 @xmath35 0.31 myr , considering the uncertainties in age estimation .
our sed results ( table [ table_sed ] ) also point out that barring 6 sources ( all being h@xmath7 emission line sources ; possibly associated with the evolved h ii region ) , the other 22 sources have ages less than 1.07 myr ( majority of them are located in the dark cloud at the western border of ) . the protostellar source detected by @xcite , and the cool , compact sources detected by @xcite also show recent star formation occurring in the dark cloud .
the dynamical age of the outflow for uyso1 , which was constrained by @xcite as @xmath22 10@xmath71 yr , and its location at the immediate interface of supports our notion of triggered star formation by the expanding h ii region . to examine
if the collect - and - collapse due to the ionizing star is possibly at work in the region , we compare the observed properties with the analytical model by ( * ? ? ?
* section 5 ) for star formation at the periphery of an expanding h ii region .
the model gives the expressions for the time when the fragmentation starts ( t@xmath131 ) , and the radius at the time of fragmentation ( r@xmath131 ) .
the value of @xmath132 ( isothermal sound speed in the shocked layer ) , which varies from 0.2 to 0.6 km s@xmath25 , was taken to be 0.2 km s@xmath25 for our calculations . using the values of @xmath72 and @xmath77
calculated in previous section ( section [ subsection_physical_properties ] ) , we obtained @xmath133 @xmath21 1.27 myr and @xmath134 @xmath21 1.70 pc .
this would suggest that fragmentation would start at a radius of 1.70 pc , whereas the stellar sources found are within a 1.70 pc radius of the massive star .
however , the distance of the ysos is the projected distance and the actual distance could be larger . taking the present mean age of the ysos as @xmath4 1 myr and the fragmentation time as calculated
, it would imply that the age difference between the ysos and ( @xmath4 5 @xmath35 3 myr ) is @xmath135 2 myr .
though this @xmath133 calculation suggests that this h ii region could be old enough to lead to star formation by collect - and - collapse process , the morphology of the region makes this process unlikely .
moreover , we have assumed the lowest value of isothermal sound speed for our calculations ( 0.2 km s@xmath25 ) . a higher value of @xmath132 will only further increase the fragmentation time and radius and make a collect - and - collapse scenario unlikelier .
also , the ysos are not exclusively associated with the pah emission that we see around this region ( figure [ fig_co3to2 ] ) .
pah regions represent the compressed neutral matter between the ionization and shock fronts .
the ysos extend well beyond the pah emission , as is evident from our msx 8.28 @xmath49 m contours ( which , in addition , contains continuum emission too ) .
this is distinctly separate from the star forming regions of , for example , @xcite where young sources and condensations have been observed along the pah ring .
it should be noted that , in an ideal case , one would expect regularly - spaced - clumps / protostars of very similar ages distributed along the periphery of the h ii region , mostly along the material swept - up by the expanding ionization front .
this is definitely not the case here .
it therefore appears that the collect - and - collapse process due to the star hd 53623 is most likely not involved in the star formation in this region .
alternatively , it is possible that the star formation observed at the north - western border of sh2 - 297 is probably due to some other processes .
the formation of ysos in the dark cloud area could partly be due to the influence of expanding h ii region sh2 - 297 by rdi ( radiation driven implosion ) mechanism ( most likely upto the edge and its vicinity ) @xcite , and partly due to supernova induced star formation as discussed by @xcite ( the distant ysos possibly having formed by this process ) .
the ysos are distributed along the concave arc ( see section [ subsection_spatial_distribution ] ) of the molecular cloud traced by @xmath90co , and seem to have an age sequence .
it appears that the supernova might have swept the molecular material and led to its piling up along the arc .
therefore , as the material might have collected along the arc , beginning from the edge of sh2 - 297 towards the north - west , star formation could have occured at different epochs .
the plausible location of the supernova ( @xmath136 @xmath4 07@xmath1308@xmath14 , @xmath137 @xmath4 -11.2@xmath17 , * ? ? ? * ) and the derived age of the supernova remnant - 0.5 myr @xcite - lend weight to this hypothesis . to better understand the star formation and the effect of the massive star on the region , we need much deeper , wider and longer wavelength observations for better statistics ( so as to constrain mass and luminosity functions of the embedded ysos ) , and to carry out a study of ysos along the arc of molecular material connecting to the larger region
. it would also be helpful to carry out a radial velocity study and spectroscopic observations for a wider area to constrain the membership , age and spectral type of the ysos .
this would confirm their relative evolutionary status , identify possible clusters , and help put the relevant triggered star formation process on a firm footing .
in this paper , we have carried out a multiwavelength study of the star - forming region .
our main conclusions are as follows : 1 .
we identified the spectral class of the central ionizing star , using optical spectroscopic data .
the prominence of he i lines in our spectra indicate a spectral type of b0v .
our optical spectroscopic results are consistent with the radio continuum data results .
2 . using the radio data for 610 , 1280 , & 1420 mhz , the physical parameters of the region were derived . taking the extent of the h ii region to be 1.6 pc ( 5@xmath6 ) , the emission measure was calculated to be 9.15 @xmath35 0.56 @xmath1 10@xmath2 @xmath3 pc , and thereby the electron density as 756 @xmath35 46 @xmath0 .
these low values of emission measure and the electron density confirm that this is indeed a classical h ii region .
the strmgren radius was calculated to be 0.051 @xmath35 0.002 pc and the dynamical timescale to be 1.07 @xmath35 0.31 myr .
3 . the grism slitless spectroscopic image was used to identify the h@xmath7 emission line stars ; nir cc diagram ( _ j - h / h - k _ ) to identify the class ii - type objects , and the nir cm diagram ( _ k / h - k _ ) to identify the red sources ( @xmath138 ) in our field .
the average extinction in the field was calculated to be 2.70 mag .
we estimated the age and mass of the ysos using our optical cm diagram ( _ v / v - i _ ) , besides using the nir mass spectrum to estimate the mass range .
age was found to be in the range 0.5 to 2 myr , with the average age being @xmath4 1 myr .
the mass was estimated to be @xmath4 0.1 to 2 m@xmath10 , with a few outliers @xmath4 3 m@xmath10 .
we studied the evolutionary status of some of the sources from @xcite , red sources , class ii - type objects , and h@xmath7 emission line stars using the sed model of @xcite .
the models predict - for most of the sources - stellar age in the range 0.05@xmath91 myr , stellar mass in the range [email protected] m@xmath10 , and disk accretion rate @xmath4 10@xmath13910@xmath127 m@xmath10yr@xmath25 .
a few h@xmath7 emission line stars , which are most evolved ysos in the field , have stellar ages upto 5 myr , and masses @xmath4 2.5@xmath95 m@xmath10 .
the spatial distribution of the ysos indicates a possible evolutionary sequence , in the sense that youngest sources are distributed away from the ionizing source .
this could be evidence in support of triggered star formation in the region .
the star formation seems to have propagated from the ionizing source towards the cloud , which is further supported by the presence of massive protostar uyso1 discovered by @xcite , and cool , compact sources discovered by @xcite , which are located away from the ionizing source .
we thank the anonymous referee for a very thorough and critical reading of our manuscript .
the useful comments and suggestions made by the referee greatly improved the scientific content of the paper .
the authors thank the staff of hct , operated by indian institute of astrophysics ( bangalore ) ; igo at girawali , operated by inter university centre for astronomy & astrophysics ( iucaa ) , pune ; irsf at south africa in joint partnership between s.a.a.o and nagoya university of japan ; and gmrt managed by national center for radio astrophysics of the tata institute of fundamental research ( tifr ) for their assistance and support during observations .
this paper used data from the nrao vla archive survey ( nvas ) .
the nvas can be accessed through http://www.aoc.nrao.edu/@xmath4vlbacald/. mt is supported by kakenhi 22000005 .
kkm acknowledges support from a marie curie irses grant ( 230843 ) under the auspices of which some part of this work was carried out .
+ 1 & 07:04:58.030 & -12:16:50.80 & & & & & & 14.12 @xmath35 0.08 & 10.78 @xmath35 0.02 & 8.62 @xmath35 0.09 + 2 & 07:05:00.665 & -12:16:45.13 & & & 16.85 @xmath35 0.10 & 14.67 @xmath35 0.05 & 12.96 @xmath35 0.04 & 11.12 @xmath35 0.01 & 9.73 @xmath35 0.01 & 7.17 @xmath35 0.02 + + 3 & 07:04:55.818 & -12:17:37.90 & & & & 16.62 @xmath35 0.03 & 15.51 @xmath35 0.09 & & & + 4 & 07:04:56.305 & -12:17:24.14 & & & & 16.82 @xmath35 0.03 & 15.47 @xmath35 0.07 & 14.78 @xmath35 0.11 & 13.69 @xmath35 0.10 & + 5 & 07:04:56.334 & -12:16:14.28 & & & & 17.32 @xmath35 0.04 & 16.18 @xmath35 0.13 & & & + 6 & 07:04:56.580 & -12:15:14.47 & & & & 16.67 @xmath35 0.02 & 15.80 @xmath35 0.10 & 14.63 @xmath35 0.08 & & + 7 & 07:04:56.634 & -12:16:59.69 & & & 17.39 @xmath35 0.04 & 14.95 @xmath35 0.01 & 13.60 @xmath35 0.02 & 12.64 @xmath35 0.03 & & + 8 & 07:04:56.898 & -12:16:39.00 & & & & 15.87 @xmath35 0.07 & 15.09 @xmath35 0.06 & 13.71 @xmath35 0.06 & 12.49 @xmath35 0.08 & + 9 & 07:04:57.702 & -12:16:20.06 & & & & 13.32 @xmath35 0.01 & 12.21 @xmath35 0.01 & 11.06 @xmath35 0.02 & 10.91 @xmath35 0.02 & + 10 & 07:04:57.927 & -12:20:36.37 & & & & 17.14 @xmath35 0.05 & 16.12 @xmath35 0.13 & & & + 11 & 07:04:58.407 & -12:17:47.23 & & & & 16.60 @xmath35 0.03 & 15.27 @xmath35 0.06 & 14.78 @xmath35 0.10 & 14.21 @xmath35 0.16 & + 12 & 07:04:58.749 & -12:16:18.35 & & & & 16.27 @xmath35 0.03 & 15.15 @xmath35 0.06 & & & + 13 & 07:04:59.150 & -12:15:17.78 & & & & 16.73 @xmath35 0.03 & 16.05 @xmath35 0.14 & & & + 14 & 07:04:59.994 & -12:15:41.34 & & & 17.65 @xmath35 0.03 & 16.12 @xmath35 0.02 & 15.34 @xmath35 0.06 & & & + 15 & 07:05:00.319 & -12:16:43.75 & & & & 15.69 @xmath35 0.03 & 13.44 @xmath35 0.02 & & & + 16 & 07:05:00.599 & -12:18:52.50 & & & 17.56 @xmath35 0.03 & 16.25 @xmath35 0.02 & 15.52 @xmath35 0.08 & & & + 17 & 07:05:00.826 & -12:18:46.47 & & & & 16.80 @xmath35 0.05 & 16.12 @xmath35 0.13 & & & + 18 & 07:05:00.881 & -12:18:26.03 & & & 17.55 @xmath35 0.05 & 16.86 @xmath35 0.04 & 16.10 @xmath35 0.14 & 14.59 @xmath35 0.10 & & + 19 & 07:05:00.919 & -12:15:42.37 & 19.65 @xmath35 0.01 & 15.29 @xmath35 0.01 & 12.47 @xmath35 0.01 & 10.83 @xmath35 0.01 & 10.07 @xmath35 0.01 & 9.61 @xmath35 0.01 & 9.47 @xmath35 0.01 & + 20 & 07:05:01.313 & -12:17:22.48 & & & & 16.83 @xmath35 0.03 & 15.45 @xmath35 0.08 & 14.72 @xmath35 0.10 & & + 21 & 07:05:01.873 & -12:17:11.83 & & & & 16.95 @xmath35 0.04 & 15.87 @xmath35 0.12 & 14.82 @xmath35 0.16 & & + 22 & 07:05:02.311 & -12:17:17.60 & & & & 16.99 @xmath35 0.03 & 16.21 @xmath35 0.13 & & & + 23 & 07:05:02.637 & -12:16:44.26 & & & & 16.29 @xmath35 0.02 & 14.99 @xmath35 0.05 & & 14.33 @xmath35 0.24 & + 24 & 07:05:02.833 & -12:16:12.72 & & & & 17.34 @xmath35 0.04 & 15.72 @xmath35 0.09 & 14.46 @xmath35 0.18 & 13.69 @xmath35 0.10 & + 25 & 07:05:03.666 & -12:18:24.20 & & & & 16.09 @xmath35 0.02 & 14.27 @xmath35 0.03 & 13.23 @xmath35 0.05 & 12.15 @xmath35 0.05 & + 26 & 07:05:03.832 & -12:18:00.40 & & & 16.64 @xmath35 0.02 & 15.10 @xmath35 0.01 & 14.35 @xmath35 0.03 & 13.12 @xmath35 0.06 & & + 27 & 07:05:04.448 & -12:17:09.32 & & & & 15.98 @xmath35 0.01 & 14.31 @xmath35 0.03 & 13.26 @xmath35 0.05 & 12.57 @xmath35 0.05 & + 28 & 07:05:05.984 & -12:17:21.69 & & & & 16.88 @xmath35 0.03 & 15.37 @xmath35 0.06 & & & + 29 & 07:05:05.995 & -12:16:45.27 & & & & 17.15 @xmath35 0.04 & 15.39 @xmath35 0.07 & 14.65 @xmath35 0.14 & & + 30 & 07:05:06.044 & -12:17:17.39 & & & & 16.91 @xmath35 0.04 & 15.65 @xmath35 0.09 & & & + 31 & 07:05:06.075 & -12:18:35.45 & & & & 16.55 @xmath35 0.03 & 14.59 @xmath35 0.03 & 13.27 @xmath35 0.09 & 12.06 @xmath35 0.07 & + 32 & 07:05:06.216 & -12:17:33.02 & & & 18.25 @xmath35 0.07 & 16.46 @xmath35 0.03 & 15.48 @xmath35 0.08 & & & + 33 & 07:05:06.735 & -12:17:52.68 & & & & 16.85 @xmath35 0.03 & 15.85 @xmath35 0.11 & & & + 34 & 07:05:07.139 & -12:19:23.87 & & & 17.59 @xmath35 0.05 & 15.48 @xmath35 0.01 & 14.34 @xmath35 0.03 & & & + 35 & 07:05:08.827 & -12:17:54.66 & & & 17.69 @xmath35 0.03 & 16.31 @xmath35 0.02 & 15.60 @xmath35 0.08 & & & + 36 & 07:05:08.890 & -12:15:34.44 & & & & 16.89 @xmath35 0.06 & 16.00 @xmath35 0.13 & 15.72 @xmath35 0.27 & & + 37 & 07:05:09.181 & -12:19:25.46 & & & 16.00 @xmath35 0.01 & 14.15 @xmath35 0.01 & 13.26 @xmath35 0.01 & & & + 38 & 07:05:09.283 & -12:17:24.01 & & & 17.70 @xmath35 0.04 & 16.31 @xmath35 0.02 & 15.63 @xmath35 0.09 & & & + 39 & 07:05:09.326 & -12:18:42.24 & & & 14.38 @xmath35 0.04 & 12.84 @xmath35 0.06 & 12.05 @xmath35 0.05 & 10.88 @xmath35 0.06 & 10.12 @xmath35 0.06 & + 40 & 07:05:09.749 & -12:20:02.05 & & & 17.45 @xmath35 0.04 & 16.22 @xmath35 0.02 & 15.52 @xmath35 0.08 & & & + 41 & 07:05:09.811 & -12:19:26.43 & & & & 16.97 @xmath35 0.04 & 16.06 @xmath35 0.14 & & & + 42 & 07:05:10.071 & -12:18:51.91 & & & & 18.02 @xmath35 0.10 & 16.19 @xmath35 0.14 & & & + 43 & 07:05:10.124 & -12:18:29.40 & & & 16.65 @xmath35 0.01 & 14.53 @xmath35 0.01 & 13.40 @xmath35 0.02 & & & + 44 & 07:05:10.358 & -12:18:09.88 & & & 17.46 @xmath35 0.03 & 16.26 @xmath35 0.02 & 15.61 @xmath35 0.09 & & & + 45 & 07:05:10.395 & -12:19:25.70 & & & & 16.64 @xmath35 0.06 & 15.66 @xmath35 0.10 & 11.20 @xmath35 0.10 & 10.85 @xmath35 0.12 & 5.12 @xmath35 0.07 + 46 & 07:05:10.481 & -12:18:30.70 & & & & 17.59 @xmath35 0.07 & 15.89 @xmath35 0.10 & & & + 47 & 07:05:11.475 & -12:19:16.58 & & & & 16.55 @xmath35 0.06 & 15.38 @xmath35 0.13 & & & + 48 & 07:05:11.946 & -12:18:48.14 & & & 15.46 @xmath35 0.01 & 13.99 @xmath35 0.01 & 13.32 @xmath35 0.01 & & & + 49 & 07:05:13.678 & -12:19:51.96 & & & 15.89 @xmath35 0.01 & 14.52 @xmath35 0.01 & 13.83 @xmath35 0.02 & 11.81 @xmath35 0.13 & 11.10 @xmath35 0.10 & + 50 & 07:05:17.596 & -12:20:23.48 & & & & 16.88 @xmath35 0.04 & 16.14 @xmath35 0.13 & & & + 51 & 07:05:17.668 & -12:20:26.39 & & & 17.47 @xmath35 0.04 & 16.66 @xmath35 0.03 & 15.86 @xmath35 0.11 & & & + 52 & 07:05:23.035 & -12:19:56.15 & & & 17.03 @xmath35 0.02 & 16.28 @xmath35 0.02 & 15.64 @xmath35 0.08 & & & + + 53 & 07:04:56.673 & -12:16:12.77 & 20.19 @xmath35 0.03 & 16.94 @xmath35 0.01 & 15.81 @xmath35 0.01 & 14.25 @xmath35 0.01 & 13.29 @xmath35 0.02 & 12.04 @xmath35 0.02 & 11.43 @xmath35 0.02 & + 54 & 07:04:57.964 & -12:17:13.85 & & & 16.08 @xmath35 0.01 & 15.07 @xmath35 0.01 & 14.37 @xmath35 0.04 & 14.14 @xmath35 0.06 & 12.92 @xmath35 0.06 & + 55 & 07:04:59.938 & -12:17:24.19 & & & 17.61 @xmath35 0.05 & 15.61 @xmath35 0.03 & 14.43 @xmath35 0.04 & 13.17 @xmath35 0.04 & 12.94 @xmath35 0.07 & + 56 & 07:05:00.844 & -12:20:06.48 & & & 17.20 @xmath35 0.04 & 16.32 @xmath35 0.02 & 15.70 @xmath35 0.09 & & & + 57 & 07:05:01.117 & -12:16:50.42 & & & 17.35 @xmath35 0.04 & 16.08 @xmath35 0.03 & 15.13 @xmath35 0.06 & & & + 58 & 07:05:01.230 & -12:19:03.69 & 20.20 @xmath35 0.03 & 16.73 @xmath35 0.01 & 14.83 @xmath35 0.01 & 13.60 @xmath35 0.01 & 12.85 @xmath35 0.01 & 11.77 @xmath35 0.02 & 11.11 @xmath35 0.02 & + 59 & 07:05:02.673 & -12:15:44.75 & & & 16.70 @xmath35 0.02 & 15.84 @xmath35 0.01 & 15.28 @xmath35 0.07 & & & + 60 & 07:05:03.364 & -12:15:53.58 & & & 17.78 @xmath35 0.04 & 16.75 @xmath35 0.03 & 16.09 @xmath35 0.14 & & & + 61 & 07:05:04.206 & -12:18:45.13 & & & 17.90 @xmath35 0.06 & 15.30 @xmath35 0.01 & 13.80 @xmath35 0.02 & 12.44 @xmath35 0.04 & 11.37 @xmath35 0.03 & + 62 & 07:05:04.644 & -12:19:10.39 & & & 15.96 @xmath35 0.02 & 14.82 @xmath35 0.01 & 14.13 @xmath35 0.02 & 13.46 @xmath35 0.13 & & + 63 & 07:05:05.166 & -12:18:19.49 & & & 17.43 @xmath35 0.04 & 15.88 @xmath35 0.02 & 14.82 @xmath35 0.04 & & & + 64 & 07:05:08.269 & -12:19:15.83 & & & 16.16 @xmath35 0.01 & 13.69 @xmath35 0.01 & 12.16 @xmath35 0.01 & 10.41 @xmath35 0.06 & 9.03 @xmath35 0.03 & + 65 & 07:05:08.771 & -12:18:51.26 & & & 16.67 @xmath35 0.02 & 15.41 @xmath35 0.02 & 14.64 @xmath35 0.04 & & & + 66 & 07:05:09.197 & -12:19:00.00 & & & 17.09 @xmath35 0.03 & 15.29 @xmath35 0.01 & 13.95 @xmath35 0.02 & 11.16 @xmath35 0.08 & 10.86 @xmath35 0.17 & + 67 & 07:05:10.638 & -12:19:15.02 & & & 17.04 @xmath35 0.02 & 15.89 @xmath35 0.03 & 15.10 @xmath35 0.06 & & & + 68 & 07:05:10.818 & -12:19:43.90 & & & 17.16 @xmath35 0.04 & 15.83 @xmath35 0.02 & 14.89 @xmath35 0.05 & & & + 69 & 07:05:11.424 & -12:19:24.36 & & & 16.93 @xmath35 0.04 & 15.13 @xmath35 0.01 & 13.76 @xmath35 0.04 & 11.04 @xmath35 0.10 & 10.45 @xmath35 0.10 & + 70 & 07:05:11.541 & -12:18:11.39 & 20.02 @xmath35 0.05 & 16.58 @xmath35 0.03 & 14.63 @xmath35 0.01 & 13.09 @xmath35 0.01 & 12.15 @xmath35 0.01 & & 10.04 @xmath35 0.08 & + 71 & 07:05:11.642 & -12:19:30.95 & & & 15.10 @xmath35 0.01 & 13.96 @xmath35 0.01 & 13.14 @xmath35 0.01 & & & + 72 & 07:05:11.774 & -12:16:24.01 & 20.78 @xmath35 0.05 & 17.95 @xmath35 0.02 & 15.34 @xmath35 0.01 & 13.72 @xmath35 0.01 & 12.52 @xmath35 0.01 & 11.27 @xmath35 0.02 & 10.29 @xmath35 0.02 & 7.89 @xmath35 0.07 + 73 & 07:05:14.495 & -12:20:47.57 & & & 15.25 @xmath35 0.01 & 14.16 @xmath35 0.01 & 13.47 @xmath35 0.02 & 12.43 @xmath35 0.12 & 11.65 @xmath35 0.12 & + 74 & 07:05:14.610 & -12:19:42.29 & & & 15.07 @xmath35 0.01 & 13.62 @xmath35 0.01 & 12.74 @xmath35 0.01 & & & + 75 & 07:05:15.324 & -12:20:02.05 & & & 17.17 @xmath35 0.04 & 15.89 @xmath35 0.01 & 14.95 @xmath35 0.05 & 11.61 @xmath35 0.22 & 11.30 @xmath35 0.27 & + 76 & 07:05:17.419 & -12:18:58.69 & & & 14.42 @xmath35 0.01 & 13.49 @xmath35 0.01 & 12.88 @xmath35 0.01 & 11.95 @xmath35 0.04 & 11.21 @xmath35 0.06 & + 77 & 07:05:18.113 & -12:19:25.47 & & & 17.25 @xmath35 0.03 & 15.90 @xmath35 0.01 & 14.94 @xmath35 0.05 & & & + 78 & 07:05:18.950 & -12:20:13.53 & & & 16.03 @xmath35 0.05 & 14.93 @xmath35 0.03 & 13.97 @xmath35 0.03 & & 9.47 @xmath35 0.05 & + + 79 & 07:05:06.330 & -12:15:36.38 & 17.43 @xmath35 0.01 & 14.82 @xmath35 0.01 & 12.95 @xmath35 0.01 & 12.01 @xmath35 0.01 & 11.33 @xmath35 0.01 & 10.44 @xmath35 0.01 & 9.84 @xmath35 0.01 & 7.36 @xmath35 0.03 + 80 & 07:05:09.652 & -12:19:56.32 & 20.07 @xmath35 0.03 & 16.18 @xmath35 0.01 & 14.05 @xmath35 0.01 & 12.77 @xmath35 0.01 & 12.02 @xmath35 0.01 & 10.91 @xmath35 0.07 & 10.21 @xmath35 0.07 & + 81 & 07:05:10.920 & -12:15:22.46 & & & 15.11 @xmath35 0.01 & 14.49 @xmath35 0.01 & 14.11 @xmath35 0.03 & 13.61 @xmath35 0.04 & 13.12 @xmath35 0.05 & + 82 & 07:05:12.292 & -12:18:38.27 & 16.74 @xmath35 0.01 & 14.37 @xmath35 0.01 & 13.01 @xmath35 0.01 & 12.06 @xmath35 0.01 & 11.50 @xmath35 0.01 & 11.05 @xmath35 0.11 & 10.53 @xmath35 0.12 & + 83 & 07:05:15.359 & -12:19:39.02 & & & 13.13 @xmath35 0.06 & 12.27 @xmath35 0.06 & 11.99 @xmath35 0.05 & 10.93 @xmath35 0.12 & 10.27 @xmath35 0.10 & + 84 & 07:05:15.840 & -12:19:45.09 & 15.41 @xmath35 0.14 & 14.39 @xmath35 0.02 & 13.41 @xmath35 0.01 & 12.55 @xmath35 0.01 & 12.10 @xmath35 0.01 & & & + 85 & 07:05:16.265 & -12:20:17.61 & & & 14.32 @xmath35 0.01 & 13.37 @xmath35 0.01 & 12.84 @xmath35 0.01 & & & + 86 & 07:05:17.904 & -12:15:40.26 & 15.27 @xmath35 0.01 & 14.17 @xmath35 0.01 & 13.89 @xmath35 0.01 & 13.65 @xmath35 0.02 & 13.40 @xmath35 0.03 & 12.86 @xmath35 0.03 & 12.50 @xmath35 0.04 & + 87 & 07:05:18.494 & -12:15:09.22 & 17.35 @xmath35 0.01 & 14.85 @xmath35 0.01 & 14.00 @xmath35 0.01 & 12.85 @xmath35 0.01 & 11.95 @xmath35 0.01 & 10.77 @xmath35 0.01 & 9.97 @xmath35 0.01 & 8.21 @xmath35 0.03 + 88 & 07:05:19.389 & -12:20:28.71 & 14.37 @xmath35 0.01 & 12.92 @xmath35 0.01 & 11.83 @xmath35 0.01 & 11.11 @xmath35 0.01 & 10.79 @xmath35 0.01 & 10.09 @xmath35 0.06 & 9.50 @xmath35 0.05 & + 89 & 07:05:19.592 & -12:17:31.48 & 16.82 @xmath35 0.01 & 14.59 @xmath35 0.01 & 13.38 @xmath35 0.01 & 12.51 @xmath35 0.01 & 12.10 @xmath35 0.01 & 11.40 @xmath35 0.01 & 10.74 @xmath35 0.02 & 7.82 @xmath35 0.03 + 90 & 07:05:20.068 & -12:19:12.59 & 13.86 @xmath35 0.01 & 12.69 @xmath35 0.01 & 11.67 @xmath35 0.01 & 10.92 @xmath35 0.01 & 10.40 @xmath35 0.01 & 9.82 @xmath35 0.01 & 9.36 @xmath35 0.01 & 6.62 @xmath35 0.04 + 91 & 07:05:20.728 & -12:21:42.69 & 17.11 @xmath35 0.01 & 15.14 @xmath35 0.01 & 13.83 @xmath35 0.01 & 12.84 @xmath35 0.01 & 12.15 @xmath35 0.01 & 11.03 @xmath35 0.02 & 10.38 @xmath35 0.02 & + 92 & 07:05:21.912 & -12:19:07.19 & 19.80 @xmath35 0.02 & 15.92 @xmath35 0.01 & 14.30 @xmath35 0.01 & 13.44 @xmath35 0.01 & 13.01 @xmath35 0.01 & 12.53 @xmath35 0.03 & 12.02 @xmath35 0.03 & + 93 & 07:05:22.359 & -12:17:18.63 & 18.23 @xmath35 0.01 & 15.39 @xmath35 0.01 & 14.17 @xmath35 0.01 & 13.30 @xmath35 0.01 & 12.84 @xmath35 0.01 & 12.25 @xmath35 0.02 & 11.62 @xmath35 0.03 & + + 94 & 07:05:12.139 & -12:14:41.57 & & & 14.15 @xmath35 0.04 & 13.26 @xmath35 0.04 & 12.77 @xmath35 0.03 & 12.26 @xmath35 0.02 & 11.76 @xmath35 0.02 & 9.38 @xmath35 0.03 + 95 & 07:05:16.154 & -12:14:55.90 & 18.57 @xmath35 0.01 & 15.73 @xmath35 0.01 & 14.24 @xmath35 0.03 & 13.30 @xmath35 0.03 & 12.77 @xmath35 0.03 & 12.15 @xmath35 0.02 & 11.53 @xmath35 0.02 & 10.00 @xmath35 0.09 + 96 & 07:05:23.669 & -12:22:36.08 & 16.18 @xmath35 0.01 & 14.14 @xmath35 0.01 & 13.25 @xmath35 0.03 & 12.48 @xmath35 0.02 & 12.16 @xmath35 0.02 & 11.89 @xmath35 0.02 & 11.46 @xmath35 0.03 & 8.33 @xmath35 0.10 + 97 & 07:05:37.457 & -12:21:17.32 & & & 13.98 @xmath35 0.02 & 13.14 @xmath35 0.02 & 12.73 @xmath35 0.03 & 12.04 @xmath35 0.05 & 11.34 @xmath35 0.07 & + | we present a multiwavelength study of the galactic h ii region , located in canis major ob1 complex .
optical spectroscopic observations are used to constrain the spectral type of ionizing star as b0v .
the classical nature of this h ii region is affirmed by the low values of electron density and emission measure , which are calculated to be 756 @xmath0 and 9.15@xmath110@xmath2 @xmath3 pc using the radio continuum observations at 610 and 1280 mhz , and vla archival data at 1420 mhz . to understand local star formation
, we identified the young stellar object ( yso ) candidates in a region of area @xmath4 [email protected]@xmath6 centered on using grism slitless spectroscopy ( to identify the h@xmath7 emission line stars ) , and near infrared ( nir ) observations .
nir yso candidates are further classified into various evolutionary stages using color - color ( cc ) and color - magnitude ( cm ) diagrams , giving 50 red sources ( @xmath8 ) and 26 class ii - like sources .
the mass and age range of the ysos are estimated to be @xmath4 0.1@xmath92 m@xmath10 and 0.5@xmath92 myr using optical ( _ v / v - i _ ) and nir ( _ j / j - h _ ) cm diagrams .
the mean age of the ysos is found to be @xmath4 1 myr , which is of the order of dynamical age of 1.07 myr of the h ii region . using the estimated range of visual extinction ( 1.1@xmath925 mag ) from literature and nir data for the region , spectral energy distribution ( sed ) models
have been implemented for selected ysos which show masses and ages to be consistent with estimated values .
the spatial distribution of ysos shows an evolutionary sequence , suggesting triggered star formation in the region .
the star formation seems to have propagated from the ionizing star towards the cold dark cloud located west of .
= 1 |
there has been a recent influx of ideas and tools from quantum information theory that are being used to characterize many - body condensed matter systems .
part of this interest stems from the fact that some subtle properties of the underlying phase of matter , otherwise difficult to calculate , can be found quite naturally by partitioning the density matrix of a ground - state wave function and calculating measures of entanglement .
for example , in one dimensional critical systems , it was shown that the central charge can be determined via the scaling of the entanglement entropy which takes the form , @xmath2 where @xmath3 is the size of the partitioned region and @xmath4 is the central charge characterizing the critical point@xcite . in higher dimensional critical systems
general results remain unknown , though interesting behavior has been found in a few isolated critical systems.@xcite entanglement finds a natural place in the study of disordered systems where the concepts of entanglement and de - localization go hand in hand . for example , the entanglement entropy at infinite - randomness fixed points displays a similar @xmath5 scaling behavior in which , by analogy , it was proposed that these models may have very unconventional values for the central charge@xcite .
it was also conjectured that while the entanglement entropy has a similar behavior as the translationally invariant case , the rnyi entropies are very different , and in fact constant .
recent studies of the so - called _ single - particle _ entanglement entropy in disordered systems , in contrast to the many - body entanglement entropy described above , have found an interesting scaling behavior , @xmath6 , where @xmath3 is the system size and @xmath7 is related to the multifractal spectrum through , @xmath8 where @xmath9 are the scaling exponents of the inverse participation ratios ( iprs ) ( defined below)@xcite . in this work
we extend this result and study the rnyi entropy for non - interacting disordered systems . whereas references only examined the @xmath10 moment of the rnyi entropy @xmath0
, we show that the other moments yield more information about the multifractal spectrum .
indeed more than the derivative of the multifractal spectrum at @xmath10 can be found by looking at other moments .
specifically , for small @xmath11 the scaling behavior of the rnyi entropy and the iprs are identical .
since the multifractal spectrum for the critical wavefunction is universal at anderson transition , the rnyi entropy for small @xmath11 is also universal @xcite .
thus we show that as in the case of critical one dimensional systems , universal information can indeed be found through entanglement .
the multifractal spectrum ( mfs)@xcite , like the central charge for critical 1 + 1 dimensional systems , plays an important role in characterizing non - interacting disordered systems@xcite .
more concretely , the mfs is the set of scaling exponents of the ipr , defined as @xmath12 which is indexed by a real number @xmath13 it has been observed that in many critical disordered systems the iprs have a spectrum of scaling exponents defined by @xmath14 characterized by the spatial dimension @xmath15 and function @xmath16 which is the mfs .
generically , the mfs can be written as @xmath17 where @xmath18 is a function of @xmath11 that depends on the system .
for a conventional band insulator @xmath19 while in a metal @xmath20 . at a critical point , the @xmath18 can have a non - trivial dependence on @xmath11@xcite .
the mfs and its symmetry properties play a central role in restricting the possible types of critical field theories one can construct@xcite .
an understanding of the multifractal nature of critical wave functions is hence important for an understanding of the localization transition .
additionally , advances in imaging have also made it possible to begin to observe the multifractal spectrum in a number of disordered systems at the anderson transition@xcite .
our interest , however , in the connection between the mfs and the rnyi entropies , is chiefly theoretical . while for the sake of argument we consider models where the multifractal spectra are known , in general this set of exponents is difficult to compute .
we demonstrate that a much simpler calculation gives an accurate approximation to an interesting region of the mfs . by an explicit calculation and two numerical examples ,
we show that the information content of critical wavefunctions of disordered free - fermion models contains imprinted signatures of their mfs .
the entanglement measures we focus on are the standard von - neumann entanglement entropy and a closely related quantity , the rnyi entropy . given a wave function @xmath21 , it is possible to construct the density matrix as , @xmath22 .
the conventional entanglement entropy is calculated by partitioning the system into two spatial subregions @xmath23 and @xmath24 , where only one subregion is observed , say region @xmath23 .
one then traces out all of the possible configurations of region @xmath24 to yield the reduced density matrix @xmath25 the question is then asked how much information does @xmath23 know about region @xmath24 and an obvious candidate for such a measure is the von - neumann entropy of region @xmath23 , @xmath26 _ i.e. _ the entanglement entropy .
a closely related quantity is the set of rnyi entropies , @xmath27parameterized by a real number @xmath11 such that as @xmath28 the rnyi entropy reduces to the von - neumann entropy .
we are focusing on localization in terms of single particle physics and thus we will use a notion of single - particle entanglement for a choice of a single - particle eigenstate .
one can define entanglement using a site occupation number basis in the second quantized fock space@xcite . from now on , we work with a single - particle eigenstate @xmath29 which can be written in the single particle occupation basis as done by previous authors@xcite . explicitly , the wave function can be written as @xmath30 where @xmath31 is the normalized single particle wavefunction and @xmath32 is the state with @xmath11 particles on site @xmath33 .
the wavefunction can then be written as , @xmath34 where @xmath35 are states where the single particle is located in region @xmath23 at site @xmath36 .
the matrix @xmath37 generically has the form , @xmath38 @xmath39 denoting the amplitude of the wavefunction at site @xmath40 in region @xmath41 respectively .
the schmidt decomposition of such a wavefunction is easily carried out .
using singular value decomposition , we can define a new basis @xmath42 and @xmath43 where the matrices @xmath44 are the unitary matrix of singular vectors for the matrix @xmath45 .
that is , they satisfy the relationship @xmath46 where @xmath47 is the diagonal matrix containing the singular values of the matrix @xmath37 .
there are only two non - zero elements in the single particle problem : @xmath48 and @xmath49 .
the reduced density matrix @xmath50 is then easily written since the density matrix is now diagonal .
one simply finds that @xmath51^n+\left[\sum_{i\in b}b_{i}^2\right]^n = p_{a}^n+p_{b}^n\]]where @xmath52 is the probability that a particle in the single - body wavefunction @xmath53 lies in region a or b respectively .
hence , we may write the rnyi entropy as @xmath54 which agrees with the expression for the entanglement entropy obtained by ref . as @xmath55 since
we have restricted our attention to the one particle sector of the density matrix and not the full many - body density matrix , the entanglement entropy should not grow large with the subsystem size as has been noted previously@xcite .
the expectation is that the same is true here .
all the interesting behavior is in the scaling with the total system size . indeed , by looking at the entanglement of a single site
, we can extract the dominant scaling behavior with the system size , @xmath3 .
looking at the case where region @xmath23 is a single site @xmath56 , we simply have @xmath57 where @xmath15 is the dimension . clearly , any site @xmath56 is not especially privileged and hence , one should look at the average over all possible sites , chosen with a uniform probability distribution . denote @xmath58_{x_{i } }
$ ] as this site average with respect to the uniform probability distribution .
that is , in @xmath15 dimensions each site @xmath56 is chosen with equal probability @xmath59 where @xmath60 is the lattice constant .
if we look at _ the site averaged _
@xmath61 we obtain , @xmath62_{x_{i } } & = \left ( \frac{a}{l } \right)^{d } \ \sum_{x_{i } } \left ( ( 1-p_{a}(x_i))^{n } + p_{a}^{n}(x_i ) \right ) \nonumber \\ & = & \left ( \frac{a}{l } \right)^{d } \left [ \sum_{m=0}^{\infty } \left ( \begin{array}{c } n \\ m \end{array } \right ) ( -1)^{m } \sum_{x_{i } } p_{a}^{m } + \sum_{x_{i } } p_{a}^{n } \right ] \nonumber \\ & = & \left ( \frac{a}{l } \right)^{d } \left [ \sum_{m=0}^{\infty } \left ( \begin{array}{c } n\\m \end{array } \right ) ( -1)^{m } p_m+p_n\right ] \nonumber\\ & \sim & \left ( \frac{a}{l } \right)^{d } \left [ \sum_{m=0}^{\infty } \left ( \begin{array}{c } n\\m \end{array } \right ) ( -1)^{m } \left(\frac{a}{l}\right)^{\tau(m)}+\left(\frac{a}{l}\right)^{\tau(n)}\right ] \nonumber\\\end{aligned}\ ] ] the quantity @xmath63 can be identified with the inverse participation ratio , @xmath64 which we recall scales as @xmath65 , where the @xmath16 are the multifractal spectrum .
one finds that for @xmath66 the multifractal spectrum is strictly less than zero and the site - averaged @xmath67 can be approximated as , @xmath68_{x_{i } } \simeq 1 - n(a / l)^{d } + \left(a / l\right)^{\tau(n)+d}.\ ] ] by extension , the site averaged rnyi entropy can be approximated as , @xmath69 which is the main result of this work and clearly shows the relationship between the rnyi entropies and the multifractal spectrum .
grassberger@xcite found a related connection between rnyi entropies and the multifractal spectrum of the probability measures of classical strange attractors .
as @xmath70 , as expected , our result reproduces the result of jia et .
al @xcite for the von neumann entropy , up to a subleading correction that vanishes as @xmath71 .
indeed , there is always a region where @xmath16 is less than zero where our approximation is valid since the multifractal spectrum is a non - decreasing , convex function with @xmath72 , the spatial dimension , and @xmath73 . for moments @xmath74 ,
such a simplification can not be achieved and according to eq.(8 ) , we need to sum up all the terms to get the rnyi entropies .
the connection between the moments of the rnyi entropies and the multifractal spectra for these large moments is less clear .
we will now consider two well - studied models where the multi - fractal spectrum is known and compare our approximate form with the known results .
we first consider a system of dirac fermions confined to 2d in the presence of a quenched random magnetic field normal to the plane with vanishing magnetic flux on average .
such a system can be thought of as the continuum limit of various tight binding hamiltonians like the chalker - coddington network model@xcite .
general considerations have predicted that the random magnetic field dirac model has an exact zero - mode critical wave function which is multi - fractal . as has been noted by numerous authors ,
the most useful property of this model is that the wavefunction can be calculated _ exactly _ for any realization of the random magnetic field @xmath75 and is given by @xmath76 where @xmath77 is the normalization factor . here
the random magnetic field @xmath78 is assumed to be gaussian distributed with variance ( disorder strength ) @xmath79 , _ i.e. _ @xmath80 \propto \exp[-\frac{1}{2 g } \int d^{2}x [ \nabla\phi]^{2 } ] $ ] @xcite . to construct the reduced density matrix we simply note
@xmath81 and the multifractal spectrum @xmath82 for the random - flux dirac model for system sizes l=[100,1000 ] . in figures
( a ) , the system is in the strong disorder regime while in figures ( b),(c),(d ) the system is in the weak disorder regime . in both , we find good agreement between the scaling of the rnyi entropy and multifractal spectrum for exponents @xmath83.[fig2 : ] ] depending on the value of @xmath79 , its been shown that the multifractal spectrum for a dirac fermion in a random magnetic field has two regimes . in the weak disorder regime ,
defined by @xmath84 , the multifractal spectrum displays a discontinuous behavior as a function of @xmath11 that was found in ref .
[ ] : @xmath85 in the strong disorder regime , where @xmath86 , the similarity between the rnyi entropy and the multifractal spectrum can also be seen for values of the exponents given by @xmath87 in both the weak and strong disorder regimes eq . [ eqn : result ] is applicable as long as @xmath88 in figure [ fig2 : ] , we plot a few values of @xmath11 and compare the scaling behavior of the rnyi entropy and the multifractal spectrum in the strong disorder and weak disorder regimes for system sizes @xmath89 $ ]
. good agreement between the two quantities is found . in the prbm .
the scaling behavior is shown for @xmath90 and @xmath91 for system sizes l=[100 , 1000 ] . in figures
( a ) and ( b ) , the systems is in the weak multifractal regime with @xmath92 and @xmath93 respectively . in figures ( c ) and ( d ) ,
where @xmath94 and @xmath95 respectively , we can see that the agreement is good in the strong multifractal regime . in figure ( d ) , only @xmath96 and @xmath97 are shown for the sake of clarity . ]
to see that eq .
[ eqn : result ] is quite general for any single particle wave function , we explicitly check our result in a much simpler model : the periodic random banded model ( prbm ) .
the model is defined as the ensemble of random hermitian @xmath98 matrices where the entires @xmath99 are independently distributed gaussian variables with mean zero and a variance that falls off as @xmath100 at @xmath101 the model undergoes an anderson transition from the localized @xmath102 to delocalized @xmath103 phases , for all values of @xmath104 . at this point ,
the model shows key features of the anderson critical point , namely eigenfunctions possessing multifractal behavior .
again there is a single extended wave function on which we will focus .
we expect and confirm that the previous analysis holds for moments less than @xmath88 however , unlike the previous case , an analytic expression for the wave function is not known so we match our approximation to the numerical calculation of ipr scaling . to eliminate the effect of boundaries , we will look at the periodic version of this model where @xmath105 this can be interpreted as describing a 1d model with long range hopping that falls off as @xmath106 @xcite . from our previous results , it is expected that the combination @xmath107 should scale with the system size to some power , @xmath108 , where @xmath109 is predicted to be @xmath110 ( note the change from @xmath111 to @xmath110 when compared with the random flux case due to the change in spatial dimension ) .
systems of size l=[100 , 1000 ] were considered . as shown in fig .
[ fig1 : ] , @xmath109 matches very well with @xmath110 . in both the strong multifractal regime and weak multifractal regimes there is a range of values for which the scaling of the rnyi entropy and the inverse participation ratio are nearly identical .
in conclusion , we showed that looking at the extended wave functions in the single particle occupation basis yields a general relationship between the rnyi entropies and multifractal spectrum .
our results can be applied to a wide - range of models , two of which were explicitly shown in this work .
we hope that the rnyi entropy formula offers a simpler and more efficient way to calculate universal pieces of the multifractal spectrum .
the connection also opens the door to measuring rnyi entropies experimentally through the measurement of the multifractal spectrum via imaging .
finally , it would be interesting to explore to what extent the rnyi entropy and the multifractal spectrum are related for many - body ground state wave functions . _ acknowledgments _ this work was supported in part by the national science foundation through the grants dmr 0758462 ( ef , tlh ) and dmr-1064319 ( ef ) at the university of illinois .
bh was supported by nsf grant phy-1005429 . | we show that the rnyi entropies of single particle , critical wave functions for disordered systems contain information about the multifractal spectrum .
it is shown for moments of the rnyi entropy , @xmath0 , where @xmath1 , it is possible to extract universal information about the multifractility of such systems .
this is shown through a generic calculation and then illustrated through two example models .
we find good agreement between our analytic formula and numerical simulations of the two test models .
our formalism is easily extendable to generic non - interacting fermion models .
it is also suggested that recent experimental advances in measuring the multifractal spectrum might allow some moments of the rnyi entropy to be measured . |
observations of the temperature anisotropies of the cosmic microwave background ( cmb ) , particularly those from the wilkinson microwave anisotropy probe ( wmap ) @xcite , form the foundations of the remarkably successful `` concordance '' cosmological model @xcite .
an essential ingredient of this model is the assumption that the primordial density fluctuations that seeded the formation of galaxies and large - scale structure were statistically homogeneous and gaussian .
versions of the inflation scenario based on the idea of a single slow - rolling scalar field predict levels of non - gaussianity too small to be observed . on the other hand ,
multi - field inflation models , and models with a non - standard kinetic term for the inflaton , may yield larger non - gaussian effects which could in principle be detected in current or next - generation observations @xcite .
analysis of currently available wmap data provide strong limits on the level of non - gaussianity @xcite . on the other hand , @xcite recently reported a detection of primordial non - gaussianity at greater than 99.5% significance .
further detailed analyses of non - gaussianity are clearly necessary in order to reconcile and understand the various constraints and claimed detections .
the greatest barrier to the detection of non - gaussianity , or other departures from the framework of the concordance cosmological model , is the presence of residual foreground contamination or other systematic errors .
since our own galaxy emits at microwave frequencies , the emission from local foregrounds must be carefully cleaned before a map can be obtained that is suitable for analysis .
one way of avoiding this problem is to cut out regions near the galactic plane where contamination is particularly severe , but this throws away the advantage of having full coverage .
it is therefore important to produce maps that are as clean as possible over the whole sky for many purposes .
however , such cleaning is inevitably approximate and biases are bound to occur @xcite .
circumstantial evidence exists that may be interpreted as being due to the presence of residual galactic foregrounds in the wmap data , or some other artefact of the cleaning process @xcite , but in the absence of a more complete characterization of the galactic emission the situation remains unclear . in this
_ article _ we propose and test a simple diagnostic analysis that offers the possibility of identifying foreground - related biases and systematics in all - sky maps of the cmb .
the statistical variation of the cmb temperature , @xmath4 , over the celestial sphere can be conveniently decomposed into spherical harmonic modes : @xmath5 where the @xmath6 are spherical harmonic functions , defined in terms of the legendre polynomials , @xmath7 , using @xmath8 and the @xmath9 are complex coefficients which can be expressed with @xmath10 where @xmath11 are the phases @xcite .
we have adopted the condon - shortly phase definition in the spherical harmonic decomposition .
the spherical harmonics functions , @xmath12 , can be visualized by considering their _ nodal lines _ , i.e. the set of points @xmath13 on the sphere where the spherical harmonics vanish , i.e. where @xmath14 .
nodal lines are circles , which are either in the latitude or longitude direction with respect to the coordinate system being used ; in the case of cmb maps this is usually the galactic coordinate system .
the number of nodal lines of each type is determined by the number of zeros of @xmath12 in the latitudinal and longitudinal directions independently .
the associated legendre functions @xmath7 possess @xmath15 zeros in the latitude direction , whereas the trigonometric @xmath16 and @xmath17 functions possess @xmath18 zeros in the longitude direction .
two specific orders @xmath1 are of particular interest at a given @xmath0 in the context of this work : these denote the _ zonal _ modes , with @xmath3 , and the _ sectoral _ modes , with @xmath19 . in the former case
there are no zero - crossings in the longitude direction , so contours of equal temperature run parallel to latitude lines ; in the latter the contours run parallel to longitude lines .
examples are shown in figure 1 . in the intermediate cases with @xmath20
there are zero - crossings in both directions , giving rise to a patchwork appearance ; these are usually called _ tesseral _ modes . statistically isotropic gaussian random cmb temperature fluctuations on a sphere , of the type that result from the simplest versions of the inflationary paradigm , possess spherical harmonic coefficients ( @xmath9 ) whose real and imaginary parts are mutually independent and both gaussian @xcite .
the statistical properties of the fluctuations are then completely specified by the angular power spectrum , @xmath21 , where @xmath22 since @xmath23 is always real , the complex vectors of the @xmath9 on the argand plane for @xmath24 are mirror images of those with @xmath25 with respect to @xmath26 axis for even @xmath1 , and with respect to @xmath27 axis for odd @xmath1 .
writing @xmath28 we note that the variances of the real and imaginary parts of @xmath9 for @xmath25 are equal @xmath29 which depends only on @xmath30 .
the distributions of @xmath31 and @xmath32 are independent gaussians with these variances and zero mean .
the amplitudes @xmath33 for these modes therefore have a rayleigh distribution with random phases @xcite . for @xmath3 ,
the imaginary part of @xmath9 must be zero , so this mode always has zero phase @xmath34 .
this is because the phase relates to the variation around the polar axis only .
however , the distribution @xmath31 for @xmath3 should be equal to that of the other real parts at a given @xmath30 , namely a gaussian with zero mean and stated variance .
statistical analysis of the spherical harmonics of cmb maps usually involves using all modes in an equivalent manner .
however , since the zonal modes in particular have such a special relationship to the galactic coordinate system , it is worth looking at their properties independently to see whether they hold any clues to possible residual contamination aligned with the galactic plane . in order to construct a test which involved the smallest possible number of assumptions , and in particular avoided the need to make estimates of the power - spectrum @xmath21 along the way , we focused on the modes with maximum or minimum amplitude at a given @xmath30 .
if all modes are statistically equivalent then the different orders @xmath1 at a given @xmath30 are equally likely to furnish the maximum ( or minimum ) amplitude @xmath33 . a preference for modes with @xmath3 to display the maximum ( or minimum ) amplitude
might therefore be plausibly interpreted as evidence that the zonal modes are either contaminated with residual foreground , or that foregrounds have been excessively subtracted .
both of these possibilities are supported by other evidence @xcite .
the maps analyzed in this paper are the 1-year @xcite , 3-year @xcite , and 5-year @xcite internal linear combination ( ilc ) maps from the wmap team , the 1 and 3 year maps from @xcite and the harmonic ilc map by @xcite ( hereafter ilc1 , ilc3 , ilc5 , toh1 , toh3 and hilc ) .
we apply three tests to these maps , which are described in detail below . for our first test
, we analyze the value of @xmath1 associated with the largest ( or smallest ) amplitude @xmath33 for a given @xmath30 in the ranges @xmath35 $ ] and @xmath36 $ ] .
since the @xmath9 for all orders @xmath1 at a given @xmath30 should have the same variance in the null hypothesis ( of a stationary gaussian random field ) , the minimum or maximum value of @xmath33 should not occur preferentially at any particular value of @xmath1 .
we use a very straightforward method to establish whether this is the case .
for each map we simply count the number of occurrences ( i.e. numbers of separate degrees @xmath30 within the range analyzed ) for which the minimum or maximum value amplitude occurs at @xmath3 ( the zonal mode ) or @xmath19 ( the sectoral mode ) .
these counts are recorded in table [ tabnum ] .
what we have done is to identify , at each degree @xmath30 , the value of @xmath1 which has the minimum ( or maximum ) value of @xmath33 .
if the minimum is at @xmath3 then this contributes to the count in column 3 of the table ; if the maximum is @xmath3 it contributes to the count in column 4 ; likewise occurrences of extrema at @xmath19 contribute to the counts in the following two columns . for results where the maximum @xmath30 = 10 the counts are out of a total of 11 and for maximum @xmath30 = 20 they are out of a total of 21 .
.this table shows , for the various cleaned all - sky maps described in the text , the maximum value of the degree @xmath30 considered , and , in the following four columns , the number of times ( i.e. number of values of @xmath30 ) for which the minimum ( or maximum ) amplitude @xmath9 is at @xmath3 ( zonal mode ) or @xmath19 ( sectoral mode ) . for example , the @xmath37 and max @xmath30 = 20 result for toh3 means that 11 out of 21 different values of the degree @xmath38 $ ] have the lowest value of the amplitude @xmath33 at @xmath3 . [ cols="^,^,^,^,^,^",options="header " , ] the trend reinforced again in this table seems to be that the anomalous result becomes increasingly significant as the maps supposedly improve .
we have presented a simple statistical analysis based on properties of the zonal modes of cosmic microwave background maps , i.e. those aligned parallel to the galactic plane .
an application of the test to various cleaned cmb maps gives interesting results .
at the 95 per cent level , no significant anomalies appear in the wmap ilc maps @xcite but there seems to be a significant tendency in some other maps @xcite to have zonal modes with systematically lower amplitudes than would be expected in the concordance model .
intriguingly , the maps that provide the most significant departures from the behavior expected under the null hypothesis are those based on later issues of the wmap data .
we have also showed that the low variance of the maps is more significant where you consider only the zonal amplitudes .
this shows that zonal modes have a notable contribution towards this low variance . finally , we considered the distribution of maxima and minima in zonal maps which again reinforced the earlier finding that the anomalies increase with later data releases
. the question of whether this is due to increasing over subtraction , or whether decreasing noise is revealing an underlying anomaly such as a systematic problem , can not be distinguished with current data . of course
the maps themselves are _ _ not statistically independent . indeed ,
if the cleaning processes involved were perfect then they would all be identical .
the different results we have found for the different maps are attributable to the nature of the zonal modes and their extra sensitivity to the structures associated with the galactic plane .
the appearance of significant anomalies in some maps rather than others is not a statistical fluke but is clear evidence that some cleaning methods leave artifacts in the distribution of mode amplitudes .
it must be noted that the probabilities we quote of around 98 to 99 percent are not overwhelming , so the results we have obtained are indicative rather than decisive .
this is not surprising , given the relatively small number of modes we used .
however , we repeated the analysis for a coordinate system aligned with the ecliptic , rather than galactic plane , and found no significant results at all .
this lends further credence to our interpretation of the outcome of our analysis in terms of an effect related to over - subtraction of galactic emission @xcite , which becomes increasingly pronounced with each data release .
a more definitive result will have to wait until more detailed foreground subtraction can be attempted , such as will be the case with the _ planck _ satellite .
jo short receives support from a science & technology facilities council ( stfc ) doctoral training grant .
we would like to thank pavel naselsky for interesting discussions .
we gratefully acknowledge the use of the healpix package @xcite .
99 alishahiha m. , silverstein e. , tong d. , 2004 , phys . rev .
d. , 70 , 123505 arkani - hamed n. , creminelli p. , mukohyama s. , zaldarriaga m. , 2004 , jcap , 4 , 1 bartolo n. , matarrese s. , riotto a. , 2002 , phys .
d. , 65 , 103505 bartolo n. , komatsu e. , matarrese s. , riotto a. , 2004 , phys .
rept . , 402 , 103 battefeld d. , battefeld t. , 2007 , jcap , 5 , 12 bennett c. l. et al . , 2003 , apjs , 148 , 1 bernardeau f. , uzan j .-
, 2002 , phys .
d. , 66 , 103506 bond j. r. , efstathiou g. , 1987 , mnras , 226 , 655 chen x. , easther r. , lim e. a. , 2007 , jcap , 6 , 23 chiang l .- y . , naselsky p. d. , coles p. , 2004
, apj , 602 , l1 chiang l .- y .
, naselsky p. d. , coles p. , 2007 ,
apj , 664 , 8 chiang l .- y
. , naselsky p. d. , coles p. , 2009 ,
apj , 694 , 339 chiang l .- y . , coles p. , naselsky p. d. , olesen p. , 2007
, jcap , 1 , 21 coles p. , 2005 ,
nat , 433 , 248 coles p. , dineen p. , earl j. , wright d. , 2004 , mnras , 350 , 989 creminelli p. , senatore l. , zaldarriaga m. , & tegmark m. , 2007 , jcap , 3 , 5 dvali g. , gruzinov a. , zaldarriaga m. , 2004 , phys .
d. , 69 , 083505 eriksen h. k. et al .
, 2005 , astro - ph/0508196 grski k. m. , hivon e. , banday a. j. , wandelt b. d. , hansen f. k. , reinecke e. m. , bartelmann m. , 2005 , apj , 622 , 759 hikage c. , matsubara t. , coles p. , liguori m. , hansen h. k. , matarrese s. , 2008 , mnras , 389 , 1439 hinshaw g. et al . , 2009 , apjs , 180 , 225 jarosik n. et al .
, 2007 , apjs , 170 , 263 kim j. et al
. , 2008 , phys .
d , 77 , 103002 komatsu e. et al .
, 2003 , apjs , 148 , 119 koyama k. , mizuno s. , vernizzi f. , wands d. , 2007 , jcap , 11 , 24 larson d. l. , wandelt b. d. , 2004 , apj , 613 , l85 - 88 lyth d. h. , ungarelli c. , wands d. , 2003 , phys .
d. , 67 , 023503 monteserin c. et al .
, 2008 , mnras , 387 , 209 naselsky p. d. , verkhodanov o. v. , nielsen m. t. b. , 2008 , astbu , 63 , 216 spergel d. n. et al . , 2007 , apjs , 170 , 377 stannard a. , coles p. , 2005 ,
mnras , 364 , 929 tegmark m. et al . , 2003 , phys .
d , 68 , 123523 yadav a. p. s. , wandelt b. d. , 2008 , phys .
lett . , 100 , 181301 | all - sky maps of the cosmic microwave background temperature fluctuations are usually represented by a spherical harmonic decomposition , involving modes labeled by their degree @xmath0 and order @xmath1 ( where @xmath2 ) .
the _ zonal modes _ ( i.e those with @xmath3 ) are of particular interest because they vary only with galactic latitude ; any anomalous behavior in them might therefore be an indication of erroneous foreground substraction . we perform a simple statistical analysis of the modes with low @xmath0 for sky maps derived via different cleaning procedures from the wilkinson microwave anisotropy probe ( wmap ) , and show that the zonal modes provide a useful diagnostic of possible systematics .
cosmology : cosmic microwave background cosmology : observations methods : data analysis |
clusters of galaxies are used for various kinds of cosmological studies .
clusters can be observed out to large distances and hence they can trace the distribution of matter on large scales . furthermore , a comparison of the properties of these distant clusters with that of nearby clusters reveals the evolution that clusters undergo between a redshift of 0.5 - 1 and now .
both large - scale structure and cluster evolution depend sensitively on cosmological parameters .
another interesting point is that clusters are closed systems : no matter can leave the deep potential well .
therefore all the metals that have been processed inside a cluster must still be present , i.e. all the traces of the formation process of the cluster and its galaxies are still observable .
clusters the largest bound structures in the universe for many applications can be regarded as being representative of the universe as a whole , because they accumulate matter from a relatively large volume ( a few tens of mpc ) .
moreover , the crossing time (= the time it takes a galaxy to move from one end of the cluster to the other ) is not much smaller than a hubble time .
that means , not all of the traces of the formation process are wiped out , but some information about the early universe is still present . in the following
we will present selected topics which are very promising for cosmology : the distribution of dark and baryonic matter , and a brief overview of various ways to study structure formation and the evolution of different cluster components by investigating their interaction . throughout this article we use @xmath1 km / s / mpc .
= 10.2 cm .2 in
the deep potential wells of galaxy clusters have accumulated matter from a large volume .
hence the ratio of baryonic to total matter in clusters is representative of the universe as a whole if all the matter is accumulated indiscriminately .
most of the baryons are in the intra - cluster medium ( icm ) the hot gas that fills all the space between the galaxies ( see fig . 1 ) .
typical values for the gas mass fraction are 10 - 25% , while the mass fraction in galaxies is only 3 - 5% of the total cluster mass .
measuring distributions and ratios of baryonic and non - baryonic matter in galaxy clusters is particularly interesting because there are several independent methods to determine the total cluster mass . in this article
i will concentrate on the mass determination by x - ray observations . assuming that the gas traces the potential well (= hydrostatic equilibrium )
the total mass can be measured through the x - ray surface brightness distribution and the gas temperature .
this approach is correct only if the magnetic pressure is negligible compared to the thermal pressure .
otherwise the mass is underestimated . to test the influence of the magnetic fields on the mass determination we used magneto - hydrodynamic simulations by dolag et al .
we found that the magnetic field causes only an underestimation of the total mass in the very centre of relaxed clusters and even there it is only a very small effect of typically not more than 5% ( dolag & schindler 2000 ) .
however , if the clusters are not relaxed , but in the process of merging , the mass can be underestimated considerably . therefore the mass determination should be either restricted to virialised clusters or be performed very cautiously in clusters which are not in equilibrium .
several groups determined gas mass fractions from x - ray observations in samples of nearby and distant clusters , e.g. mohr et al . ( 1999 ) : @xmath2 , ettori & fabian ( 1999 ) : @xmath3 , arnaud & evrard ( 1999 ) : @xmath4 , schindler ( 1999 ) : @xmath5 ( see also fig . 2 ) .
all these determinations depend on the radius where the mass fraction is determined , because the gas mass fraction increases slightly with radius . therefore the gas mass fractions have to be calculated for equivalent radii . in the above
mentioned analyses the mass was determined within a radius @xmath6 from the cluster centre .
this radius encompasses a volume that has a density of 500 @xmath7 the critical density of the universe @xmath8 . out to this radius the x - ray profiles necessary for the analysis could be measured reliably .
= 8.2 cm .2 in to determine @xmath0 , the gas mass fraction @xmath9 must be compared to the baryon density in the universe @xmath10 .
burles & tytler ( 1998a , b ) found @xmath11 from primordial nucleosynthesis .
the ratio of the baryon density and the gas mass fraction yields an upper limit for the matter density @xmath12 .
the baryon fraction can also be determined in a different way : measurements of the sunyaev - zeldovich effect inverse - compton scattering of the cosmic microwave background photons by the hot intra - cluster gas shifts the cmbr spectrum to slightly higher energies . as this effect is proportional to the gas density
, the density profile can be determined directly .
only an additional measurement of the gas temperature is necessary from x - rays .
the mean gas mass fraction found by grego et al .
( 2001 ) in a sample of 18 clusters ( @xmath13 ) is very similar to the x - ray results .
hence they derive also a similar upper limit for the matter density @xmath14 . in these analyses only the mass in the intra - cluster gas was taken into account .
baryons in the galaxies were neglected .
if they were to be included , the baryon fraction would increase slightly and hence ever more stringent constraints on @xmath0 could be placed .
for cosmology and cluster formation it is very interesting to know how the baryon fraction in clusters changes with time . in fig .
[ fgas_z ] the gas mass fraction is plotted versus the redshift .
it is clear that with the previous rosat and asca observations it was very hard to see any evolution due to the large observational uncertainties . with xmm and chandra
we will be able to measure gas mass fractions out to redshifts of almost unity with good accuracy . even with the old observations one can see clearly that the gas fraction varies considerably from cluster to cluster .
these different baryon fractions reflect probably the very early distribution of baryonic and non - baryonic matter and they are therefore of high interest for cosmology .
= 8.2 cm .2 in
clusters of galaxies can be used to study how structures form on large scales .
the formation and evolution of clusters depends very sensitively on cosmological parameters like the mean matter density in the universe @xmath0 ( thomas et al .
1998 ; jenkins et al . 1998 ; beisbart et al . 2001 ) .
therefore it is of great importance to determine the dynamical state of clusters at different redshifts , i.e. at different evolutionary states .
there are different approaches to determine the dynamical state .
x - ray observations yield a projected image of the square of the icm density and hence they often give a first idea . a much more detailed picture
can be obtained through the temperature distribution of the icm , because shocks and gas compressions due to mergers are visible clearly in temperature maps .
the excellent capabilities for spatially resolved spectroscopy of the new x - ray observatories xmm and chandra finally make a detailed temperature analysis possible , see e.g. the temperature maps by arnaud et al .
( 2001 ) , neumann et al .
( 2001 ) , markevitch et al .
( 2000 ) .
optical observations of the clusters galaxies are also very useful for the determination of the evolutionary state , because both the spatial distribution and the velocity distribution of the cluster galaxies are good indicators . in the shocks produced during the merging process particles
can be accelerated to relativistic energies .
these relativistic particles and hence a previous merger event can be detected in two ways .
( 1 ) inverse compton scattering of photons from the cosmic microwave background radiation by these relativistic particles can cause a hard x - ray excess in the spectrum , which has been discovered in a few clusters ( fusco - femiano et al .
1999 , 2000 , 2001 ) .
( 2 ) due to the presence of magnetic fields , synchrotron radiation is emitted in the form of radio halos ( markevitch & vikhlinin 2001 , feretti et al .
2001 ) . for a detailed analysis
it is necessary to know the distribution of the magnetic field within the cluster . in a recent analysis we found that the magnetic field decreases with radius almost in the same way as the icm density decreases with radius ( dolag et al .
2001 ) .
hence , the optimal approach for the determination of the dynamical state is a combination of many wavelengths x - ray , optical , radio and hard x - ray .
the various components in a cluster interact with each other . in particular the interaction between the cluster galaxies and
the icm is very important .
hence this interaction is more and more studied at the different wavelengths .
not only the energy budget and entropy of the icm are influenced but also the icm content of heavy elements .
the amount of iron in the icm is e.g. of the same order of magnitude as that in the cluster galaxies ( mushotzky 1999 ) .
the heavy elements in the icm can not be of primordial origin , but must have been produced in the cluster galaxies and subsequently been transported into the icm .
several transport processes are possible : ram - pressure stripping , galactic winds , galaxy - galaxy interaction , and jets from active galaxies .
so far very controversial results have been obtained in various studies which tried to determine for example the dominant metal enrichment process at a given evolutionary state ( see e.g. metzler & evrard 1994 ; gnedin 1998 ; murakami & babul 1999 , aguirre et al .
if we can find the correct answers , we will learn a lot about galaxy formation and cluster formation .
= 5.2 cm .2 in the new x - ray satellites xmm and chandra revolutionise also this field .
it is now possible to measure reliable metallicities out to a redshift of about unity , which was hardly possible before with e.g. asca measurements ( schindler 1999 ) .
furthermore , the metal distribution within the clusters can be measured with much higher accuracy than up to now with e.g. bepposax observations ( de grandi & molendi 2001 ) .
also different chemical elements can be distinguished now . in order to interpret these excellent new data
correctly several numerical simulations have been performed to study the different enrichment processes , for example , ram - pressure stripping of galaxies when they move through the icm ( abadi et al .
1999 , quilis et al .
2000 , mori & burkert 2000 , schulz & struck 2001 ; toniazzo & schindler 2001 ; see also fig .
[ fig : toniazzo1 ] ) .
comprehensive simulations on cluster scales will give us new insights into the various metal enrichment processes of the icm .
= 8.2 cm .2 in = 6.8 cm .2 in
observations of galaxy clusters with the new x - ray satellites xmm and chandra have revealed several surprising results .
the high spatial resolution of chandra enables us to look in detail at the cluster centres . in some clusters
`` depressions '' in the x - ray emission have been found , e.g. rbs797 ( schindler et al . 2001 , see also fig .
[ fig : rbs797_cen ] and fig .
[ fig : traces ] ) , perseus cluster ( fabian et al . 2000 ) , and hydra - a ( mcnamara et al . 2000 ) . in these clusters
the intra - cluster gas has been pushed from the low x - ray emission regions to the high x - ray emission regions by the pressure of relativistic particles in radio jets , which come from the active galaxy in the cluster centre .
these clusters are hence ideal objects to investigate the interaction of jets with the hot intra - cluster medium . in several chandra observations
surprising `` edges '' have been discovered , e.g. in a3667 ( vikhlinin et al .
these jumps in the x - ray surfaces brightness are not shock fronts , because the temperature increases while the surface brightness drops , i.e. the two sides are in pressure equilibrium .
therefore they were named `` cold fronts '' .
they have a width of about 5kpc , which is 2 - 3 times smaller than the coulomb mean free path .
it is likely that the transport processes here are suppressed by magnetic fields .
the xmm spectra of cooling flow regions in the centres of some clusters revealed that cool gas below a temperature of 1.5kev is absent , e.g. in a1835 ( peterson et al . 2001 ) .
this absence of cool gas was very surprising as it provides a serious problem to standard cooling flow models .
clusters of galaxies are ideal diagnostic tools to determine cosmological parameters in very different ways . in this article only a small selection is presented .
the determination of cluster masses and baryon fractions in clusters of galaxies is one of the very active fields at the moment .
also the formation process of clusters is studied in different ways .
interaction of different cluster components ( galaxies , intra - cluster medium ) is a topic of increasing interest , because the interaction leaves traces in the metallicities which can be used to infer details of the formation process of clusters and of galaxies .
in particular the new x - ray telescopes xmm and chandra yield at the moment extremely exciting results which give new insights in cluster physics and cosmology .
these observations will soon answer many of the open questions in cosmology .
markevitch , m. , vikhlinin , a. , mazzotta , p. , vanspeybroeck , l. , 2000 , in proc .
`` x - ray astronomy 2000 '' ( palermo , september 2000 ) , r. giacconi , l. stella , s. serio ( eds . ) , asp conference series , astro - ph/0012215 * discussion * 0.4 cm * n. panagia : * about the anti - correlation found between x - ray and radio emission , i am wondering which is the cause and which the effect : is it the radio structure that pushes the x - ray emitting gas or , viceversa , is it the hot gas that confines the radio emitting matieral ?
0.3 cm * g. kanbach : * radio observations show that cosmic ray pressure could influence the intra - cluster medium . is the cosmic ray pressure generally taken into account in the dynamical balance of the intra - cluster medium ?
0.3 cm * s. schindler : * so far cosmic ray pressure is usually not taken into account .
for standard mass determinations from x - ray observations only the thermal pressure of the intra - cluster medium is taken into account .
0.3 cm * s. schindler : * different types of initial magnetic seed fields were used : completely homogeneous and also chaotic initial magnetic field structures .
this field is amplified by compression during the cluster collapse .
it was shown that the final field structure is determined only by the dynamics of the cluster collapse and not by the inital conditions . | various cosmological applications of galaxy clusters are presented .
clusters are used to determine the baryon fraction , dark matter distribution and the matter density @xmath0 of the universe .
they also contain a wealth of information about structure formation and evolution on different scales : large - scale structure , cluster formation and also galaxy formation via the interaction of cluster galaxies with the intra - cluster medium .
in particular , the x - ray satellites chandra and xmm yield exciting new results for galaxy cluster physics and cosmology . # 1 # 2 _ mem . soc . astron .
it . _ * # 1 * , # 2 # 1 # 2 _ the messenger _ * # 1 * , # 2 # 1 # 2 _ astron .
nach . _ * # 1 * , # 2 # 1 # 2 _ astron .
astrophys .
_ * # 1 * , # 2 # 1 # 2 _ astron .
astrophys .
lett . _ * # 1 * , l#2 # 1 # 2 _ astron .
astrophys .
rev . _ * # 1 * , # 2 # 1 # 2 _ astron .
astrophys .
suppl .
ser .
_ * # 1 * , # 2 # 1 # 2 _ astron .
j. _ * # 1 * , # 2 # 1 # 2 _ ann .
rev .
astron .
astrophys . _ * # 1 * , # 2 # 1 # 2 _ astrophys . j. _ * # 1 * , # 2 # 1 # 2 _ astrophys .. j. lett .
_ * # 1 * , l#2 # 1 # 2 _ astrophys .
j. suppl .
_ * # 1 * , # 2 # 1 # 2 _ astrophys .
space sci .
_ * # 1 * , # 2 # 1 # 2 _ adv .
space res .
_ * # 1 * , # 2 # 1 # 2 _ bull .
astron .
inst .
czechosl . _ * # 1 * , # 2 # 1 # 2 _ j. quant .
spectrosc .
. transfer _ * # 1 * , # 2 # 1 # 2 _ mon . not .
r. astr .
soc . _ * # 1 * , # 2 # 1 # 2 _ mem .
r. astr .
soc . _ * # 1 * , # 2 # 1 # 2 _ phys .
lett .
rev . _ * # 1 * , # 2 # 1 # 2 _ publ .
astron .
. japan _ * # 1 * , # 2 # 1 # 2 _ publ .
astr .
soc .
pacific _ * # 1 * , # 2 # 1 # 2 _ nature _ * # 1 * , # 2 epsf.sty |
one of the biggest challenges in our understanding of qcd is the nature of colour singlet exchange or _ diffractive _ interactions .
the electron - proton collider hera is an ideal place to study hard diffractive processes in deep - inelastic @xmath0 scattering ( dis ) . in such interactions ,
the point - like virtual photon probes the structure of colour singlet exchange , similarly to inclusive dis probing proton structure .
l0.5 at hera , around @xmath1 of low @xmath2 events are diffractive @xcite .
experimentally , such events are identified by either tagging the elastically scattered proton in _ roman pot _
spectrometers @xmath3 downstream from the interaction point or by asking for a large rapidity gap without particle production between the central hadronic system and the proton beam direction . a diagram of diffractive dis is shown in fig . [ feyn ] .
a virtual photon coupling to the beam electron interacts diffractively with the proton through the exchange of a colour singlet and produces a hadronic system @xmath4 with mass @xmath5 in the final state .
if the 4-momenta of the incoming ( outgoing ) electron and proton are labeled @xmath6 ( @xmath7 ) and @xmath8 ( @xmath9 ) respectively , the following kinematic variables can be defined : @xmath10 , the photon virtuality ; @xmath11 , the longitudinal momentum fraction of the struck quark relative to the diffractive exchange ; @xmath12 , the fractional proton momentum taken by the diffractive exchange and @xmath13 , the 4-momentum squared transferred at the proton vertex .
bjorken - x is given by @xmath14 .
for the measurements presented here typical values of @xmath15 are @xmath16 .
@xmath17 denotes the inelasticity , where @xmath18 is the @xmath0 cms energy . a diffractive _ reduced cross section _
@xmath19 can be defined via @xmath20 which is related to the diffractive structure functions @xmath21 and the longitudinal @xmath22 by @xmath23 except at the highest @xmath24 , @xmath25 to a very good approximation .
if the outgoing proton is not detected , the measurements are integrated over @xmath26 : @xmath27 .
the proof @xcite that qcd hard scattering factorization is valid for diffractive dis justifies the expression of @xmath19 , at fixed @xmath15 and @xmath26 , as a convolution of diffractive proton parton distributions ( dpdf s ) @xmath28 and partonic cross sections @xmath29 : @xmath30 the @xmath28 should obey the dglap evolution equations and the @xmath31 are the same as for standard dis . in consequence , the framework of nlo qcd can be applied to diffractive dis in a similar way as to inclusive dis .
an extra assumption which is often made is that the @xmath32 dependence of @xmath33 can be factored out ( regge factorization ) into a _ pomeron flux factor _ @xmath34 : @xmath35 which is often parameterised using regge phenomenology as @xmath36 where @xmath37 is the pomeron trajectory with intercept @xmath38 .
the @xmath39 are _ pomeron pdfs_. although there is no firm basis in qcd for eq .
[ eq : reggefact ] , it is approximately consistent with present data .
the h1 collaboration has performed several new measurements of diffractive dis , presented in the form of the reduced cross section @xmath40 , which are summarised in fig .
[ fig : stamph1 ] . two high precision measurements in the kinematic range @xmath41 ( labeled ` h1 99 ' @xcite and ` h1 97 ' @xcite )
are based on the rapidity gap method . with respect to previous h1 data @xcite , a factor of
typically 5 more data were analysed and the kinematic range was extended significantly to lower @xmath42 and @xmath43 .
another new measurement using a forward proton spectrometer covers @xmath44 ( ` h1 99 - 00 fps ' , @xcite ) .
good agreement between the different data sets and measurement techniques is observed .
the zeus collaboration has presented new data @xcite on @xmath45 for @xmath46 , shown in fig .
[ fig : stampzeus ] .
diffractive events are selected by reconstructing a low value of @xmath5 in the detector .
the data correspond to a factor 2 increase in integrated luminosity with respect to previous results @xcite .
zeus has also measured ( fig .
[ fig : zeusloq2 ] , @xcite ) the diffractive cross section in the transition region @xmath47 between dis and quasi - real photon - proton interactions ( @xmath48 ) . also shown in fig .
[ fig : zeusloq2 ] is a fit based on a dipole model @xcite in which the photon fluctuates into @xmath49 and @xmath50 states .
both collaborations have measured the @xmath26 dependence of the cross section using their forward proton spectrometers . in the measured range @xmath52 ,
an exponential dependence @xmath53 is observed and the slope parameter @xmath54 is determined for different values of @xmath15 ( fig .
[ fig : bslopes ] ) .
models based on regge phenomenology predict an increasing steepness of the @xmath26 dependence with energy ( _ shrinkage _ ) : @xmath55 . as can be seen in fig .
[ fig : bslopes ] , the data at low @xmath15 are so far inconclusive .
zeus has studied @xcite the cross section dependence on the azimuthal angle @xmath51 between the electron and proton scattering planes , which is sensitive to the longitudinal cross section @xmath56 via the interference between transverse and longitudinal photon induced contributions , which introduces an asymmetry in @xmath51 .
the measured asymmetries are small and , within the present statistical uncertainties , compatible with zero . by making fits to the @xmath15 dependence of the diffractive dis data using eqs .
[ eq : reggefact],[eq : reggeflux ] , values for the effective pomeron intercept @xmath38 are determined .
results for measurements at different @xmath42 are shown in fig .
[ fig : alpha ] and compared with the _ soft pomeron _ value of 1.08 as well as with a parameterisation based on inclusive dis data , obtained from fits of the form @xmath57 where @xmath58 .
the diffractive data suggest an increase of @xmath38 with @xmath42 and are significantly higher than the soft pomeron value for @xmath59
. however , the naive expectation that the diffractive cross section grows twice as fast with energy as the inclusive one , corresponding to @xmath60 , is hard to reconcile with the present data at high @xmath42 .
h1 has studied @xcite the logarithmic @xmath42 dependence of the ratio of the diffractive to the inclusive dis reduced cross sections , presented as the quantity @xmath61 in fig . [
fig : ratio ] for different @xmath43 and @xmath15 values .
taking out the factorising @xmath15 dependence , the same behaviour is observed at different @xmath15 : at low @xmath43 , the logarithmic @xmath42 dependences of @xmath33 and @xmath62 are very similar , indicating that the ratio of the diffractive to the total proton gluon density is approximately constant in this region .
by contrast , at the highest @xmath63 , the logarithmic @xmath42 derivative becomes negative , suggestive of the presence of @xmath42-suppressed higher twist contributions to @xmath33 for @xmath64 .
however , this behaviour can also be explained by dglap evolution due to the the kinematic limit @xmath65 ( @xmath66 ) for gluon radiation in the case of @xmath33 .
h1 has presented @xcite a new nlo dglap qcd fit to the diffractive reduced cross section data , from which diffractive parton densities are determined .
the shape of the dpdf s is assumed to be independent of @xmath15 ( regge factorization ) , in agreement with the data , as can be seen in figs .
[ fig : betadep.flux ] and [ fig : q2dep.flux ] .
the precisely measured @xmath43 and @xmath42 dependences of the data are compared with the fit result , dividing out the @xmath15 dependence , parameterised as stated in eq .
[ eq : reggeflux ] .
the diffractive exchange is parameterised by a light flavour singlet and a gluon distribution at a starting scale @xmath67 and evolved to higher @xmath42 using the standard nlo dglap equations .
for the first time in diffraction , the experimental and model uncertainties are propagated to obtain error bands for the dpdf s .
the result of the fit is presented in fig .
[ fig : h1pdfs ] .
the diffractive pdf s remain large up to large fractional momenta @xmath68 ( or @xmath43 ) and are dominated by the gluon distribution . in total @xmath69 of the exchange momentum is carried by gluons at @xmath70 . to test factorization ,
the dpdf s can be used for updated comparisons with diffractive final state data from hera and the tevatron @xcite .
several new high precision data sets of inclusive diffractive dis have recently become available . in the case of the rapidity gap method ,
the data are now systematically limited at low @xmath42 .
these data are very useful to constrain the free parameters of qcd motivated models ( e.g. dipole models ) for diffractive dis and to distinguish between different approaches .
in particular , they have been used for the determination of a new generation of diffractive parton distributions including their uncertainties in the framework of nlo qcd , enabling tests of qcd factorization as applied to diffraction by comparisons with final state cross sections . | new precision measurements of inclusive diffractive deep - inelastic @xmath0 scattering interactions , performed by the h1 and zeus collaborations at the hera collider , are discussed .
a new set of diffractive parton distributions , determined from recent high precision h1 data , is presented . |
optimal transport problems , also known as monge - kantorovich problems , have been very intensively studied in the last years giving rise to numerous and important applications to pde , shape optimization and calculus of variations , so we witnessed a spectacular development of the field .
the interested reader may look at the monographs and lecture notes @xcite where the subject is fully developed .
let us briefly recall the formulations of the monge - kantorovich problems .
let @xmath0 be two compact metric spaces and let @xmath1 be a borel cost function .
the monge problem is formulated as follows : given two probability measures @xmath2 , @xmath3 find a measurable map @xmath4 such that @xmath5 ( @xmath6 denotes the push - forward of measures ) and such that @xmath7 minimizes the total cost , i.e. @xmath8 it may happens that the set of admissible maps is empty ( e.g. @xmath9 and @xmath10 ) .
then the problem could be reformulated in its kantorovich s relaxation : find @xmath11 such that @xmath12 and @xmath13 ( @xmath14 and @xmath15 are the projections on the factors of @xmath16 ) and such that @xmath17 minimizes the total cost , i.e. @xmath18 the admissible measures @xmath17 for the kantorovich problem are called transport plans .
we denote by @xmath19 the set of transport plans with marginals @xmath20 and @xmath21 .
if @xmath7 is admissible for the monge problem then the measure associated in the usual way to the graph of @xmath7 , i.e. @xmath22 , is admissible for the kantorovich problem . however the class of admissible measures for the kantorovich problem is never empty as it contains @xmath23 .
moreover , the kantorovich problem is a linear one .
existence of minimizers for the monge problem is difficult and may fails , while for the kantorovich problem the semicontinuity of @xmath24 is enough to ensure existence of minimizers .
if @xmath25 , and @xmath26 is the distance function , for @xmath27 the cost @xmath28 defines a distance on @xmath29 called @xmath30-wasserstein distance . let us recall that by kantorovich duality ( see @xcite ) the @xmath31-wasserstein distance between @xmath20 and @xmath21 , which we will simply denote by @xmath32 , can be expressed as follows@xmath33 where @xmath34 denotes the set of lipschitz function having lipschitz constant not greater than one . a relevant tool in mass transportation theory is constituted by the disintegration theorem ( theorem [ disintegration ] ) of measures which states that every transport plan @xmath35 can be written as @xmath36 where @xmath37
. we shall call _ disintegration map _ every @xmath38 such that @xmath39 in this paper we relate the structure of the set of transport plans @xmath19 with the push - forward of disintegration maps . indeed ,
given the measure @xmath40 , obviously @xmath20 is the first marginal of @xmath17 , while the second marginal depends on the disintegration map @xmath41 .
an interesting feature of transport plans appears by looking to the measure @xmath42 .
precisely , if @xmath43 is another transport plan , it results ( see lemma [ push - lemma ] ) : @xmath44 therefore , the second marginals can be fixed by looking to the push - forward of disintegration maps . in this way
the set of transport plans @xmath19 can be structured in transport classes ( see definition [ class ] ) by setting @xmath45 \leftrightarrow f_\ # \mu = g_\ # \mu$ ] . roughly speaking ( see example [ discrete - class ] )
, fixing a transport class leads to consider a constrained transport problem with respect to splitting masses or traveling ones .
lemma [ push - metric ] shows that all transport plans induced by transport maps belong to the same transport class .
moreover , by density of transport maps in @xmath46 , it follows that ( see proposition [ prop - class ] ) such transport class characterizes the transport plans induced by transport maps .
+ therefore , in this perspective the monge problem can be seen as a constrained kantorovich problem , namely @xmath47 \right \},\ ] ] for a given transport map @xmath48 . by density of transport maps , the kantorovich transport problem also corresponds to @xmath49}^w \right \},\ ] ] for a given transport map @xmath48 .
+ hence the monge problem represents a particular case of a more rich structure of problems , obtained by fixing a transport class in the kantorovich formulation of transport problems . in this context ,
considering transport problems in a fixed transport class is quite natural .
since we are looking to the push - forward trough disintegration maps , fixing a transport class results equivalent ( see section @xmath50 ) to consider measures @xmath51 satisfying the _ barycenter _ constraint @xmath52 the corresponding transport class is given by the transport plans @xmath53 such that @xmath54 . in this formulation we see that fixing a transport class is equivalent to consider transport maps @xmath41 sending @xmath20 into @xmath55 .
so it is natural to consider the following monge - kantorovich problem in the class @xmath55 : @xmath56 the above transport problem leads to consider an abstract monge problem between the space @xmath57 and @xmath58 .
let us consider the following transport cost @xmath59 we set @xmath60 for every transport class @xmath55 ( see proposition [ mklambda ] ) we have @xmath61 therefore , every existence result for the monge problem @xmath62 in the abstract setting corresponds to an existence result for the monge - kantorovich problem in the transport class @xmath55 . of course , minimizing in a transport class could be as difficult as for the monge problem since , of course , the transport classes are not in general closed .
however , by this reformulation it comes out that the monge case is peculiar .
more precisely , since the monge problem is a particular case of transportation in a transport class , one may asks what happens for others transport classes .
in other words , the matter consists in establishing if the abstract monge problem admits solutions .
the existence results for the monge problem are usually stated in the following form : under some assumption on the spaces , on the first marginal @xmath20 and on the cost @xmath63 , for every second marginal @xmath21 the monge problem admits solutions .
for the abstract monge problem @xmath64 this is not the case . for discrete measures @xmath55 ,
see section @xmath65 , for the quadratic cost it results that @xmath64 may not admit solutions . therefore , in the abstract setting , it could be also interesting to consider , under some assumption on the spaces , on the first marginal @xmath20 and on the cost @xmath63 , the question for what kind of second marginals the corresponding monge problem admits solutions . from this point of view , in some sense the monge problem is a _
lucky _ case .
let @xmath66 be two compact sets and let @xmath67 be the space of radon measures on @xmath68 .
a map @xmath69 is said to be borel , or equivalently weakly*-measurable , if for any open set @xmath70 the function @xmath71 is a real valued borel map .
equivalently , @xmath72 is a borel map if , for any borel and bounded map @xmath73 , it results that the map @xmath74 is borel .
[ disintegration ] let @xmath75 be given and let @xmath76 be the first projection map of @xmath77 , we set @xmath78
. then for @xmath79 there exists @xmath80 such that * the map @xmath81 is borel , + * @xmath82 moreover the measures @xmath83 are uniquely determined up to a negligible set with respect to @xmath20 .
let @xmath84 , as usual we will write @xmath85 , assuming that @xmath86 satisfy the condition ( i ) and ( ii ) of theorem [ disintegration ] .
obviously , the transport plan @xmath23 corresponds to the constant map @xmath87 .
let @xmath88 , be a transport map , observe that for the transport plan @xmath89 , the disintegration theorem yields @xmath90 .
therefore , the disintegration procedure for a transport plan @xmath91 produces a map @xmath92 we shall refer to such maps @xmath41 as _ disintegration maps_. for a transport map @xmath7 the corresponding disintegration map is given by @xmath93 .
of course , it is possible to look at a disintegration map as a measurable map between @xmath57 and @xmath94 .
indeed , we have the following result . [ measurable ]
a map @xmath95 is a disintegration map if and only if @xmath41 is measurable .
let @xmath95 be measurable and let @xmath96 be an open set .
observe that @xmath97 .
for a l.s.c .
function @xmath98 over @xmath68 , define @xmath99 since @xmath32 metrizes the weak * topology of measures , we have that @xmath100 is a l.s.c . map . for every @xmath101 it results @xmath102
if @xmath41 is a measurable map , it follows that the map @xmath103 is borel as composition of a l.s.c map and a measurable one .
hence @xmath41 is a disintegration map .
vice versa , observe that by the ascoli - arzel theorem the space @xmath104 is compact with respect to the uniform convergence .
fixed a countable dense subset @xmath105 , by kantorovich duality we have @xmath106 since @xmath107 is borel , we have that , for every @xmath108 , @xmath109 defined by @xmath110 is a borel map . to check that @xmath41 is a measurable map
, it is sufficient to observe that @xmath111-r , r[):=a.\ ] ] indeed , if @xmath112 we get @xmath113 , for every @xmath114 .
hence , by definition of @xmath115 , it follows that @xmath116 and then @xmath117 . on the other hand , if @xmath117 , i.e. @xmath116 , by kantorovich duality @xmath118 for every @xmath114
this implies @xmath112 .
let @xmath119 , we recall that the barycenter of a measure @xmath120 is given by @xmath121 disintegration maps naturally produce measures of the form @xmath122 on the space @xmath94 . by the following lemma , we see that this point of view is equivalent to fix the second marginal of transport plans induced by transport maps .
[ push - metric ] let @xmath123 be two given borel maps , @xmath124 and let @xmath125 defined by @xmath126 , @xmath127 .
then @xmath128 assume @xmath129 and for any @xmath130 let us consider the function @xmath100 defined in .
observe that @xmath131 .
hence we have @xmath132 @xmath133 since @xmath134 bi we get @xmath135 by the arbitrariness of @xmath98 we infer @xmath136 .
vice - versa , for every @xmath137 let us consider the function @xmath138 .
observe that @xmath130 . if @xmath139 we compute @xmath140 @xmath141 by the arbitrariness of @xmath142 we obtain @xmath129 .
let @xmath123 be two given borel maps , @xmath124 , let @xmath125 defined by @xmath126 , @xmath127 and let @xmath91 , @xmath143 .
then @xmath144 observe that the first part of the proof of the above lemma works for general transport plans @xmath91 , @xmath143 .
actually , by we get the following [ push - lemma ] let @xmath124 , @xmath125 , @xmath145 , @xmath146 be given .
then the following implication holds true @xmath147 therefore , also for transport plans , the second marginal can be fixed by fixing the push - forward of disintegration maps .
notice that in general the converse of is not true as we show in the next example .
let @xmath148 defined by @xmath149 and let @xmath150 .
let @xmath151 where @xmath152 for a given transport map @xmath153 with @xmath154 .
for every @xmath155 we have @xmath156 while @xmath157 for any @xmath158 let us consider @xmath159 .
we compute @xmath160 however , on the other hand @xmath161 .
the above arguments allow to characterize transport plans through the push forward of disintegration maps .
we introduce the following notion of transport class .
[ class ] let @xmath162 with @xmath163 , @xmath164 be given .
we shall say that @xmath17 and @xmath165 are equivalent ( by disintegration ) , in symbols @xmath166 , if @xmath167 .
+ for any given @xmath168 with @xmath164 , we shall call transport class of @xmath165 the equivalence class of transport plans given by @xmath169=\{\gamma= f(x)\otimes \mu\;\vert\;f_\#\mu= g_\#\mu\}.\ ] ] notice that in the case of discrete first marginal @xmath170 , for any disintegration map it is easily seen that @xmath171 therefore , transport classes are fixed by the range of @xmath41 .
[ discrete - class ] let @xmath172 consider the transport plan which uniquely splits the mass at @xmath173 .
this transport plan corresponds to the disintegration map @xmath174 by changing the point at which the mass is splitted , the range of the corresponding disintegration map does not change .
for instance , for the second transport plan in figure [ fig : discrete - class ] we get the following disintegration map @xmath175 it follows @xmath176 .
analogously , all the transport plans with only one splitted mass belong to the same transport class .
+ on the other hand , by changing the number of splitted masses the corresponding disintegration range is changing . indeed , by looking at figure [ fig : discrete - class2 ] , we may consider the disintegration map @xmath177
@xmath178 in such a case we have that @xmath179 on the other hand , by keeping fixed the number of splitted masses , the transport class may be changed by modifying the amount of traveling masses .
consider for instance the disintegration @xmath180 @xmath181 we get @xmath182 therefore , to fix a transport class leads to consider a _ constrained transport problem _
, with respect to splitting masses or traveling ones . by lemma [ push - metric ]
it follows that all transport plans induced by transport maps belong to the same transport class .
since the transport maps are dense in @xmath19 we can prove the following result .
[ prop - class ] let @xmath183 , be a transport map , i.e. such that @xmath184 , with @xmath20 non - atomic , and let @xmath185 .
if @xmath186 $ ] then there exists a transport map @xmath153 such that @xmath187 , i.e. the transport plan @xmath17 is induced by a transport map @xmath7 .
in particular , if @xmath36 , it results @xmath188 @xmath20-a.e .. by applying ( * ? ? ? * theorem 9.3 ) , see also @xcite , and the same argument employed in the proof of ( ( * ? ? ?
* theorem 2.1 ) ) , we find a sequence of borel maps @xmath189 such that @xmath190 and therefore @xmath191 $ ] , @xmath192 .
consider @xmath193 by the push - forward constraint we get @xmath194 let us consider @xmath195 defined by @xmath196 .
in fact , setting @xmath197 the set of dirac deltas , the function @xmath198 is lipschitz , with respect to the wasserstein distance , over @xmath199 . hence it suffices to consider any lipschitz extension of @xmath198 on the whole @xmath58 . for every @xmath200 , since @xmath201 we have @xmath202 therefore , by passing to a subsequence , we may suppose that @xmath203 is weakly convergent and let @xmath7 be the weak limit of @xmath203 .
+ let @xmath36 . by definition of weak convergence , by approximation with continuous functions , for any @xmath204
we get @xmath205 @xmath206 therefore @xmath207 . on the other hand , since @xmath208 $ ] , i.e. @xmath209 , and by we have @xmath210 then , it follows that @xmath203 strongly converges to @xmath211 , hence by ( * ? ? ?
* lemma 9.1 ) we deduce that @xmath212 with @xmath213 .
of course , the above arguments hold true as well by considering @xmath214 with @xmath215 .
this corresponds to consider the transport cost @xmath216 .
+ proposition [ prop - class ] allows to reformulate the monge transport problem as follows : @xmath217 \right \},\ ] ] for a given transport map @xmath48 .
+ by density of transport maps , the kantorovich transport problem can be seen as @xmath218}^w \right \},\]]for a given transport map @xmath7 .
therefore , the monge problem corresponds to minimize the functional @xmath219 in a fixed transport class of @xmath19 , while the kantorovich one corresponds to minimize the same functional on the whole @xmath19 .
in the previous section we have seen that the monge problem could be seen as a particular case of minimization on a transport class .
since the transport classes are determined through the push - forward of disintegration maps , they can be assigned by probability measures @xmath220 over @xmath94 .
+ actually , consider @xmath221 and @xmath222 . since @xmath223 , for every @xmath224 we have @xmath225 @xmath226 therefore , in order to define a transport class , the measure @xmath220 has to satisfy the constraint @xmath227 hence , every probability measure @xmath55 over @xmath94 satisfying defines a transport class @xmath228=\{f\otimes \mu : f_\#\mu = \lambda \}$ ] .
+ for instance , the transport class @xmath229 $ ] corresponds to the measure @xmath230 . while for a transport map @xmath7
, the transport class @xmath231 $ ] corresponds to @xmath232 . on the other hand
, the transport class in corresponds to the discrete measure @xmath233 . in this perspective ,
transport plans in the transport class @xmath55 can be seen as transport maps between @xmath20 and @xmath220 .
it is then natural to consider the monge - kantorovich problem in the class @xmath55 defined as follows @xmath234 by proposition [ prop - class ] , the monge problem corresponds to the transport class @xmath232 . observe that for a discrete transport class @xmath235 , the monge - kantorovich problem in the class @xmath55 yields the _ optimal allocation problem _ consisting in the determination of a borel subdivision @xmath236 of @xmath57 minimizing @xmath237 under the constraint @xmath238 . for existence results in this framework see section @xmath65 .
the notion of transport class leads naturally to consider an abstract monge problem between the space @xmath57 and @xmath58 .
consider the following transport cost @xmath239 we have the following [ mklambda ] for every transport class @xmath240 we have @xmath61 it suffices to observe that for any disintegration map @xmath148 such that @xmath54 , it results @xmath241 observe that by the above proof it follows that @xmath41 is a solution of @xmath62 if and only if @xmath242 is a solution of @xmath243 .
therefore , every existence result for the monge problem @xmath62 in the abstract setting corresponds to an existence result for the monge - kantorovich problem in the transport class @xmath55 .
the abstract setting has the advantage of considering a _ nice _ cost , since it is linear with respect to the second variable .
of course the disadvantage is the passage from the space @xmath244 to the space @xmath245 . observe that by ( * ? ? ?
* theorem 2.4 ) the transport classes are always non - empty provided @xmath20 is non - atomic .
[ cost - reg ] @xmath24 is continuous , l.s.c .
, caratheodory , normal iff @xmath246 is .
where we say that a measurable map @xmath63 is caratheodory ( resp .
normal ) if @xmath247 is continuous ( resp .
l.s.c . ) .
[ reg ] let @xmath248 $ ] be a borel cost function satisfying @xmath249 continuous at @xmath250 and such that @xmath251 .
we have the following 1 .
if @xmath247 is continuous then @xmath252 ( and hence @xmath253 ) is continuous .
2 . if @xmath247 is l.s.c .
then @xmath24 ( and hence @xmath253 ) is l.s.c .
let @xmath254 be continuous .
if @xmath255 on @xmath77 we compute @xmath256 @xmath257 as @xmath258 . if @xmath259 is l.s.c
considering @xmath260 by , passing to the liminf we obtain @xmath261 in general , for the existence of optimal transport plans in the kantorovich problem at least the lower semicontinuity property of the cost function is usually required .
actually , some regularity of the cost function is needed to obtain a useful duality theory or to ensure that the infimum of the kantorovich problem is equal to the infimum of the monge problem ( see for instance @xcite ) .
however , it is not hard to verify that the kantorovich problem admits solutions under more weak requirements on the cost function ( see for instance @xcite ) . for reader s convenience here we provide some details based on disintegration maps .
+ let @xmath262 be a minimizing sequence for the monge - kantorovich problem . passing to a subsequence
we may suppose that @xmath263 .
now , for any @xmath264 , we get @xmath265 by density of continuous functions , the above limit holds for @xmath266 as well . therefore ,
if the cost function has the form @xmath267 , with @xmath268 , then by it follows that @xmath269 is an optimal transport plan . arguing component - wise , the same reasonings apply to linear costs @xmath270 . for a caratheodory cost function @xmath63 , i.e. a borel map such that @xmath259 is continuous , observe that the disintegration maps @xmath271 belongs to @xmath272 , which is the dual of @xmath273 .
therefore , by passing to a subsequence we may suppose that @xmath274 , i.e. @xmath275 the above continuity property shows that @xmath276 is an optimal transport plan , provided that @xmath277 . + if @xmath63 is a normal cost , i.e. a borel measurable map such that @xmath259 is lower semicontinuous , then it can be reduced to a caratheodory cost by standard approximation procedures . for instance ( see @xcite ) , we may write @xmath278 hence , the cost @xmath279 is caratheodory and @xmath280 .
since @xmath281 passing to the limit we obtain @xmath282 passing to the limit with respect to @xmath283 we get that @xmath284 is optimal . by representation of weakly * l.s.c .
functionals ( see @xcite ) , the same reasonings apply to normal cost on @xmath245 . for a related result see also @xcite . in the following
we compare the kantorovich problem with the abstract version formulated using the transport classes .
[ comp - kant ] for every transport class @xmath285 it results @xmath286 let @xmath287 .
we compute @xmath288 where @xmath289 .
it remains to check that @xmath290 . by , for every @xmath158 we have @xmath291 @xmath292 observe that the above definition of @xmath293 can be seen as a generalized barycenter map .
indeed , we have the following [ baric ] the generalized barycenter map @xmath294 defined by @xmath295 is @xmath31-lipschitz with respect to the wasserstein distance .
first observe that if @xmath296 then @xmath297 .
fixed @xmath298 we get @xmath299 @xmath300 taking the supremum with respect to @xmath296 it results @xmath301 observe that a probability measure @xmath55 on @xmath58 defines a transport class iff its generalized barycenter is equal to @xmath21 .
[ equal - class ] if @xmath253 is normal then there exists a transport class @xmath285 such that @xmath302 let @xmath303 be a minimizing sequence for @xmath304 .
set @xmath305 . by passing to a subsequence we have that @xmath306 with respect to the weak convergence of measures .
observe that @xmath307 @xmath308 therefore , @xmath55 defines a transport class .
consider the transport plans @xmath309 . passing to a subsequence
we may also suppose that weakly @xmath310 .
since @xmath253 is normal we get @xmath311 @xmath312 the result follows by lemma [ comp - kant ] . by the above analysis the kantorovich problem over @xmath245 is essentially equivalent to the usual kantorovich s one . indeed , if @xmath253 is normal ( for instance if the cost @xmath24 satisfies the conditions of lemma [ reg ] ) by lemma [ equal - class ] we have @xmath313 for a transport class @xmath55 .
if @xmath314 is a transport plan for @xmath315 , for such transport class @xmath55 , setting @xmath316 , by it follows that @xmath317 is an optimal plan for @xmath318 . for a related relaxation procedure see @xcite
in the previous section we have seen that the monge - kantorovich problem is essentially equivalent to the transport problem on transport classes .
of course , the monge problem reveals hard to handle . for linear cost @xmath319 , which is equivalent to the quadratic cost @xmath320 because of the expansion @xmath321
, it is relatively easy to find existence and uniqueness of optimal transport maps .
we expect some advantage by considering the special form of the cost @xmath253 for the monge problem on a transport class . in order to handle with the abstract monge problem , we review here some usual tools for solving the monge problem .
the available approaches to existence and uniqueness rely more or less on two basic facts . the first one is based on the notion of @xmath24-cyclical monotonicity .
a set @xmath322 is said @xmath24-cyclical monotone if for any finite set of pairs @xmath323 and any permutation @xmath324 the following inequality holds true @xmath325 a fundamental fact in mass transportation is that the support of every optimal transport plan is a @xmath24-cyclical monotone set and every @xmath24-cyclical monotone set is contained in the @xmath24-superdifferential @xmath326 ( or contact set ) of a @xmath24-concave function @xmath198 , where @xmath327 a function @xmath198 is said to be @xmath24-concave if there exist @xmath328 such that @xmath329 for details we refer the reader for instance to @xcite .
the @xmath24-transform of @xmath198 is defined by @xmath330 .
it can be shown that @xmath331 . if one is able to show that for @xmath20 a.e . @xmath101
the @xmath24-superdifferential is single valued , then every transport plan is supported on the graph of a transport map ( see @xcite ) .
namely , there exists a unique solution of the monge problem .
the same reasonings apply as well directly to the abstract problem @xmath332 .
in other words , if the @xmath253-superdifferentials is single - valued , then the monge problem @xmath332 admits a unique solution . in this framework
the two monge problems are essentially related . indeed , suppose that the @xmath253-superdifferentials contain just one dirac delta and let @xmath198 be a @xmath24-concave function .
we set @xmath333 and @xmath334 .
it follows that @xmath335 is @xmath253-concave and @xmath336 . recalling that @xmath337 , the following implications hold true @xmath338 @xmath339 therefore @xmath340 vice - versa , suppose that @xmath24-superdifferentials are single valued and let @xmath341 be a @xmath253-concave function .
consider the @xmath24-transform @xmath342 .
+ we have @xmath343 @xmath344 for a @xmath24-concave function @xmath345 ( see for instance ( * ? ? ?
* remark 3.12 , theorem 3.10 ) ) .
it follows that @xmath346 contains just one delta .
of course this singleton condition of superdifferentials can be achieved under additional requirements on the cost function . for a differentiable cost a general condition relies in the so called _
twist ( or spence - mirrlees in economic settings ) condition _
, i.e. @xmath347 ( see @xcite ) . for a generalization of such condition , in the case of suitable geometries
see @xcite . in the case
@xmath348 with @xmath349 a riemannian manifold and for a lagrangian cost it is enough for the cost @xmath24 to satisfy the mather s shortening principle and the connectedness of the @xmath24-superdifferential , as it is shown in ( * ? ? ? * chapter 9 ) . observe that in this case the connectedness is a key property .
this property is easily satisfied if @xmath259 is linear as happens just for the cost @xmath350 .
however , general forms of the cost @xmath24 which guarantee the connectedness of the @xmath24-superdifferential are not known . in this perspective
, the consideration of the linear cost @xmath253 could be useful .
observe that for a cost @xmath24 linear with respect to the second variable , the twist condition is not in general satisfied .
consider for instance @xmath270 for possibly not invertible jacobian matrix @xmath351 . by considering the cost @xmath253
, it may happen that the @xmath253-superdifferential contains many points , which actually are probability measures of @xmath58 , although the @xmath24-superdifferential is a singleton .
however , we have the advantage that the @xmath253-superdifferential are convex sets .
consider @xmath348 and a cost @xmath24 satisfying a shortening principle .
we briefly sketch the reasonings of ( * ? ? ?
* chapter 9 ) .
suppose that the following conditions are satisfied 1 .
there exists @xmath352 with @xmath353 such that @xmath354 intersects every nontrivial continuous curve over @xmath349 .
2 . the cost @xmath24 satisfies a shortening principle .
3 . the superdifferential @xmath326 is connected . by assumption
@xmath355 it is possible to define a function @xmath356 with @xmath357 having as domain the mid point of geodesics over @xmath349 .
indeed , by definition of @xmath24-superdifferential we get @xmath358 for every @xmath359 .
it follows @xmath360 hence , the shortening principle implies that @xmath361 by the inequality it follows that @xmath362 is well defined and moreover it results a lipschitz , actually also an holder condition works as well , map . by condition @xmath363 and
@xmath364 it is possible to show that the set of points on which @xmath365 is not single valued is of null measure .
indeed , let @xmath366 . by condition @xmath364 consider a continuous curve @xmath367 lying in @xmath368 connecting @xmath369 .
therefore it is defined the non - trivial continuous curve @xmath370 .
hence , @xmath371 . by point @xmath363
it follows that @xmath372 which is a null measure set . since @xmath373
, we would like to prove that the @xmath253-superdifferential are a.e .
single valued .
the analogue condition for @xmath253 superdifferential is @xmath374 by definition , the above inequality leads to @xmath375 a shortening principle should allow to estimate the distance @xmath376 by the distance between mid points of geodesics in @xmath349 or in @xmath29 .
the main problem is to find a null set @xmath354 , due to the fact that we are now dealing with the space @xmath377 .
assume the cost function @xmath24 satisfies the twist condition .
under differentiability requirement , for instance on an open and bounded set @xmath378 , with a uniform lipschitz condition @xmath379 is injective .
it turns out that @xmath24-concave functions @xmath198 are lipschitz on @xmath380 .
indeed , it suffices to compute @xmath381 @xmath382 moreover , if @xmath198 is differentiable at @xmath383 , for every @xmath357 , we have @xmath384 passing to the limit as @xmath385 we get @xmath386 by the arbitrariness of @xmath387 it follows @xmath388 ( see also @xcite ) . therefore , if @xmath20 is absolutely continuous with respect to the lebesgue measure , then the twist condition implies that the superdifferential @xmath389 is a singleton for @xmath20-a.e .
the same reasoning applies as well directly for the cost @xmath253 .
indeed , since @xmath390 we have that @xmath253-superdifferentials contains at most one delta iff the cost @xmath24 satisfies the twist condition .
+ in the sequel we sketch a related approach to show existence and uniqueness for the monge problem .
+ assume that @xmath63 is a caratheodory function , hence by lemma [ reg ] @xmath24 is continuous .
approximate the target measure @xmath21 by a finite convex combination of dirac deltas , say by @xmath391 .
consider the problem @xmath392 .
the optimal transport plans of this approximation problem are of course extremal points of @xmath393 .
it can be shown that these extremal points , which are supported on the graph of a @xmath24-concave function @xmath394 , are of the form @xmath395 ( see section @xmath65 ) , i.e. the monge problem @xmath396 admits a unique solution . by the uniform lipschitz condition it follows that the sequence of @xmath24-concave functions @xmath397 is equi - lipschitz .
indeed , fixed @xmath398 , for @xmath399 we have @xmath400 while for @xmath401 we have @xmath402 by the ascoli - arzel theorem we may suppose that @xmath403 , uniformly on compact subsets .
observe that @xmath198 is @xmath24-concave as well . indeed , recall that a map @xmath198 is @xmath24-concave iff @xmath404 , where @xmath405 fixed @xmath406 , we find a large integer @xmath407 such that @xmath408 . since @xmath397 is @xmath24-concave we get @xmath409 by the arbitrariness of @xmath410 we obtain @xmath411 .
let @xmath412 . by passing to a subsequence
we may suppose that @xmath413 .
since @xmath400 passing to the limit as @xmath258 we get @xmath414 therefore @xmath415 .
since the @xmath24-superdifferential is a singleton , the whole sequence @xmath416 converges to @xmath417 .
we set @xmath418 .
as limit of measurable maps , @xmath7 is measurable as well .
moreover , it results @xmath419 hence @xmath5 .
since the graph @xmath420 is supported on @xmath421 it follows , see ( * ? ? ?
* th . 3.22 ) , that @xmath7 is an optimal transport map . for a related approximation procedure of the kantorovich problem
see @xcite .
+ hence , under the twist condition for the cost @xmath253 , we would have existence and uniqueness of optimal transport plans in every fixed transport class .
therefore , an interesting question is that of finding condition on the cost @xmath24 ensuring the twist condition for @xmath253 .
to treat discrete measures we sketch existence for the monge problem in this setting . for an extensive discussion of this case
we refer to @xcite .
let @xmath422 be a discrete probability measure over a metric space @xmath68 .
let @xmath17 be an optimal transport plan between @xmath20 and @xmath21 . denote by @xmath423 . by optimality , we have that @xmath424 is a @xmath24-cyclically monotone set of @xmath77 .
define @xmath425 .
let @xmath426 .
by @xmath24-cyclically monotonicity we get @xmath427 therefore , we obtain @xmath428 under the assumption of @xmath20 @xmath24-continuous , i.e. @xmath429 for every @xmath430 , we have that @xmath431 . therefore , since the set of splitting masses is given by @xmath432 , if @xmath20 is @xmath24-continuous then @xmath433 .
hence , the transport plan @xmath17 is induced by a transport map .
let @xmath55 be an atomic transport class .
optimal transport plan for @xmath434 are of the form @xmath435 ( see @xcite ) . for every index it results @xmath436 . by linearity of @xmath253
it follows @xmath437 .
therefore , the transport map @xmath438 is optimal for @xmath439 .
consider a discrete transport class given by @xmath440 .
consider the cost @xmath441 .
for @xmath442 observe that @xmath443 therefore , @xmath62 admits solution iff @xmath444 .
analogously , for cost @xmath445 , if @xmath446 for every @xmath447 , it turns out that @xmath62 admits solution iff @xmath448 for every @xmath442 . in this section
we have seen that the usual approaches to solve the monge problem give rise to some difficulties in the setting of the transport problem in a transport class .
in some sense , these methods are specific for the transport class corresponding to transport maps .
the question to establish existence in different transport classes remains open .
we have shown that also for the case of a discrete transport class the answer could be negative .
therefore , from this point of view , the monge problem corresponds to a _ lucky _
case for the fixed transport class .
this feature of transport classes also naturally leads to the following question .
the existence results for the monge problem are usually stated in the following form : under some assumption on the spaces , on the first marginal @xmath20 and on the cost @xmath63 , for every second marginal @xmath21 the monge problem admits solutions . having in mind the abstract monge problem @xmath64 , it could be also interesting to consider , under some assumption on the spaces , on the first marginal @xmath20 and on the cost @xmath63 , the question for what kind of second marginals the corresponding monge problem admits solutions .
99 p. t. abdellaoui , h. heinich , caracterisation dune solution optimale au probleme de monge - kantorovich , bull .
france , * 127 * ( 1999 ) , 429 - 443 . n. ahmad , h. k. kim , r. j. mccann , optimal transportation , topology and uniqueness , bull . math . sci .
* 1 * ( 2011 ) 13 - 32 .
l. ambrosio , lecture notes on transport problems , in mathematical aspects of evolving interfaces " .
lecture notes in mathematics * 1812 * , springer , berlin , 2003 , 1 - 52 .
l. ambrosio , n. fusco , d. pallara , functions of bounded variation and free discontinuity problems , oxford university press , new york , 2000 .
l. ambrosio , n. gigli , g. savar , gradient flows in metric spaces and in the space of probability measures , lecture notes in mathematics , eth zrich , birkhuser , 2005 .
l. ambrosio , p. tilli , analysis in metric spaces p. bernard , young measures , superposition and transport , indiana university mathematics journal * 57 * , 1 ( 2008 ) 247 - 276 .
g. carlier , a. lachapelle , a planning problem combining calculus of variations and optimal transport , appl .
* 63 * ( 2011 ) , 1 - 9 .
j. a. cuesta - albertos , a. tuero - diaz , a characterization for the solution of the monge - kantorovich mass transference problem , statist .
* 16 * ( 1993 ) , 147 - 152 .
w. gangbo , the monge transfer problem and its applications , contemp .
, * 226 * ( 1999 ) , 79 - 104
. j. gonzalez - hernandez , j. gonzalez - hernandez , extreme points of sets of randomized strategies in constrained optimization and control problems , siam j. optim . * 15 * , no . 4 ( 2005 ) 1085 - 1104
. j. gonzalez - hernandez , j. rigoberto gabriel , j. gonzalez - hernandez , on solutions to the mass transfer problem , siam j. optim .
* 17 * , no . 2 ( 2006 ) , 485 - 499 . l. granieri , optimal transport and minimizing measures , lap lambert academic publishing , 2010 . v. levin , abstract cyclical monotonicity and monge solutions for the general monge - kantorovich problem , set valued analysis * 7 * ( 1999 ) , 7 - 32 . t. maruyama , auman - perles variational problem via theory of young measures t. maruyama , an extension o the aumann - perles variational problem , proc .
japan acad .
* 55 * , ser . a ( 1979 ) , 348 - 352 .
m. mcasey , l. mou , optimal locations and the mass transport problem a. pratelli , existence of optimal transport maps and regularity of the transport density in mass transportation problems , ph.d .
thesis , scuola normale superiore , pisa , 2003 .
rachev , l. ruschendorf , mass transportation problems vol .
i , probab .
springer - verlag , new york , 1998 . c. villani , topics in mass transportation , graduate studies in mathematics * 58 * , ams , providence , ri . c. villani , optimal transport , old and new .
springer , 2009 . | by disintegration of transport plans it is introduced the notion of transport class . this allows to consider the monge problem as a particular case of the kantorovich transport problem , once a transport class is fixed .
the transport problem constrained to a fixed transport class is equivalent to an abstract monge problem over a wasserstein space of probability measures .
concerning solvability of this kind of constrained problems , it turns out that in some sense the monge problem corresponds to a _
lucky _ case
. * keywords . *
optimal mass transportation theory . monge - kantorovich problem .
calculus of variations .
shape analysis .
geometric measure theory . +
* msc 2000 . * 37j50 , 49q20 , 49q15 . |
a hallmark of strongly correlated electron systems is the competition of ground states with different kinds of order.@xcite in this context , a central set of unsettled theoretical issues concerns the character of quantum critical points ( qcps ) in metals @xcite .
such metallic qcps have been identified in several heavy fermion compounds@xcite ; evidence for quantum critical behavior has also been found in the ruthenate sr@xmath1ru@xmath2o@xmath3
@xcite , and the cuprate and iron - based superconductors .
quantum criticality is also often invoked as a possible explanation of the `` strange '' or `` bad '' metal behavior seen in a variety of such materials @xcite . to date
, there exists no satisfactory theory of metallic qcps in @xmath4 or @xmath5 spatial dimensions although a number of field theoretic approaches have been attempted @xcite . among the unsettled issues
are : * a ) * the values of critical exponents and which properties can be expressed as scaling functions involving these exponents , * b ) * the extent to which metallic qcps are generically preempted by superconducting @xcite or other forms @xcite of auxiliary order that gap out the fermi surface , * c ) * whether there exists a `` non - fermi liquid '' metal in the quantum critical regime . controlled theoretical methods , that can be used both to benchmark the field theories and for comparison with experiments are greatly needed .
determinantal quantum monte carlo ( dqmc)@xcite in particular , in cases where the fermion sign problem can be circumvented @xcite may serve this purpose .
such methods have been successfully applied in several problems in which critical bosonic fluctuations are coupled to fermions with a dirac - like dispersion @xcite . in this paper
, we report a dqmc study of a two - dimensional sign problem - free lattice model that exhibits an `` ising nematic '' qcp in a metal at finite fermion density ; in the nematic phase , the discrete lattice rotational symmetry is spontaneously broken from @xmath6 to @xmath7 .
this is a particularly relevant qcp given that nematic order @xcite and nematic quantum critical fluctuations @xcite have been observed in many of the materials mentioned above . and temperature @xmath8 . here , the other parameters entering the hamiltonian eq .
[ eq : hfhb ] are @xmath9 , @xmath10 , @xmath11 .
the solid line marks the transition temperature , @xmath12 , between the nematic and the symmetric phases .
the line extrapolates to the @xmath13 qcp at @xmath14 .
the dashed line marks @xmath15 , where the nematic susceptibility reaches 50pale ( grey ) lines show the corresponding temperatures for the case @xmath16 ( where the fermions and pseudospins are decoupled ) . in this case
, the qcp occurs at @xmath17 . ]
our simulations are limited to finite system sizes up to @xmath0 lattice sites . within our numerical accuracy
, we find evidence for a continuous nematic quantum phase transition with critical behavior that differs significantly from that of a nematic qcp in an insulator .
specifically , in the disordered phase near the qcp , where the dimensionless quantum control parameter @xmath18 , the thermodynamic ( zero frequency ) nematic correlation function ( defined in eq .
[ eq : chi1 ] ) is consistent with the following functional description : @xmath19^\gamma \label{eq : scaling}\ ] ] where @xmath8 is the temperature , @xmath20 is the wave - vector , @xmath21 is the matsubara frequency , and @xmath22 , @xmath23 , and @xmath24 are positive dimensionful constants .
we find that the exponents in this expression take values @xmath25 and @xmath26 .
( see fig .
[ fig : static ] . )
these values are in agreement with the predictions of refs .
@xcite for the behavior of the thermodynamic susceptibility near criticality .
this implies that the uniform nematic susceptibility @xmath27 has a curie - weiss form with an effective weiss temperature which varies linearly with @xmath28 , as can be seen in figs .
[ fig : phase_diagram ] & [ fig : chi qcp ] . eq .
[ eq : scaling ] is consistent with the scaling relation , @xmath29 , where @xmath30 , @xmath31 , @xmath32 . here , @xmath33 is the conventionally defined susceptibility exponent , the correlation length exponent @xmath34 , where @xmath35 is the anomalous dimension of the nematic field , and we have introduced an `` apparent dynamical exponent , '' @xmath36 .
the values of these exponents derived from our monte carlo data are given in table [ tab : exponents ] .
we also include for comparison the exponents of the standard @xmath37-dimensional ising qcp , which apply when the coupling to fermions is set to zero . despite the accuracy of the scaling analysis ,
it is important to test whether the observed behavior is characteristic of an asymptotic quantum critical scaling regime . to this end , we define a `` quadrupolar '' correlator , @xmath38 , in terms of fermion bilinears , which has the same symmetry as the nematic correlator , @xmath39 , and therefore should have the same asymptotic critical behavior . as can be seen in figs .
[ fig : comparechibeta ] , [ fig : comparechih ] , [ fig : comparechiq ] , [ fig : chi_ff ] , @xmath38 does not obey a scaling law of the form of eq .
( [ eq : scaling ] ) over a comparable range of @xmath40 as does @xmath39 . sufficiently close to criticality
, @xmath38 may be consistent with eq .
( [ eq : scaling ] ) , but the dynamic range is relatively small and the error bars on the apparent exponents are substantial . it could be that corrections to scaling are smaller in @xmath39 than in @xmath38 , or it could be that our data simply do not reflect asymptotic quantum critical scaling . in the fermionic sector
we find results consistent with strongly renormalized fermi liquid , `` marginal fermi liquid '' @xcite , or weakly non - fermi liquid behavior down to our lowest temperatures ( @xmath41 , where @xmath42 is the fermi energy ) .
in particular , at the qcp , the effective quasiparticle weight , @xmath43 , defined in terms of the single - fermion green s function in eq . , remains substantial .
however , it monotonically decreases on cooling , with downward curvature . in a fermi liquid
, @xmath43 would approach a positive limit as @xmath44 , while in a weak non - fermi liquid , it would vanish in proportion to a small power of @xmath8 .
[ fig : fermi_surface ] shows maps of the low - energy spectral weight as a function of momentum in the disordered phase , at the qcp , and in the ordered phase .
the existence@xcite of `` cold - spots '' along the zone diagonals , where the quasiparticles are relatively weakly scattered , is apparent .
finally , no superconducting transition is found down to our lowest temperatures , although the superconducting susceptibility in the s - wave channel is peaked about @xmath45 .
( see fig .
[ fig : p_of_t ] . )
| m3 cm || c | c | c | c| critical exponents & @xmath46 & @xmath33 & @xmath35 & @xmath47 + @xmath48 ising qcp & @xmath49 & @xmath50 & @xmath51 & @xmath52 + @xmath48 _ metallic _ ising nematic qcp & @xmath53 & @xmath54 & @xmath55 & @xmath56 + [ tab : exponents ] [ sec : model ] , as a function of momentum @xmath57 across the brillouin zone for three different values of the transverse field : @xmath58 ( in the nematic phase ) , @xmath59 ( near the qcp ) , and @xmath60 ( in the disordered phase ) . in the nematically ordered phase , the data is averaged over both orientations of the order parameter .
@xmath61 is proportional to the integral of the spectral function @xmath62 over an energy window of width @xmath8 [ see eq .
( [ eq : g2 ] ) ] .
the temperature is @xmath63 , and the system size is @xmath64 .
the data were taken from systems with either periodic or anti - periodic boundary conditions .
the fermi surfaces are seen as peaks in @xmath61 .
the lower right panel shows the fermi surface of the bare band structure . ] the remainder of the paper is organized as follows : in sec .
[ sec : model ] we define the lattice hamiltonian and describe its phase diagram ; in the next section , we provide evidence for the statements above regarding thermodynamic nematic and quadrupolar correlations ( [ sec : bose ] ) , dynamical nematic and quadrupolar correlations ( [ sec : dynamics ] ) , superconductivity ( [ sec : sc ] ) , and single - fermion correlations ( [ sec : fermionic ] ) ; finally , in sec . [
sec : discussion ] , we discuss various caveats concerning the interpretation of our results , and their bearing on both prior theoretical work and the interpretation of experiments .
spinful electrons reside on the sites , while the ising pseudospins live on the bonds .
the pseudospins interact with their neighbors antiferromagnetically , and are coupled to the fermion bond density . ]
our model is illustrated in fig .
[ fig : model ] , and is defined on a two - dimensional square lattice . every lattice site hosts a single ( spinful ) fermionic degree of freedom .
each link has a pseudospin-@xmath65 degree of freedom that couples to the fermion bond - density .
the system is described by the following hamiltonian : @xmath66 where @xmath67 here , @xmath68 creates a fermion on site @xmath69 with spin @xmath70 , @xmath71 denotes a pair of nearest - neighbor sites on the square lattice , @xmath72 and @xmath73 are the hopping strength and chemical potential , respectively , @xmath74 ( @xmath75 ) denote pauli matrices that act on the pseudospin that lives on the bond connecting the neighboring sites @xmath76 and @xmath69 , @xmath77 is the nearest - neighbor ising interaction between neighboring pseudospins ( here , @xmath78 denotes a pair of nearest - neighbor bonds ) , h is the strength of a transverse field that acts on the pseudospins , and @xmath79 is the dimensionless coupling strength between the pseudospin and the fermion bond density .
the pseudospins are not related to the physical spins of the electrons ; their ordering corresponds to a nematic transition .
the model ( [ eq : h ] ) should be viewed as an effective lattice model , designed to give a nematic qcp .
microscopically , the nematic degrees of freedom could represent ( via hubbard - stratonovich transformation ) interactions involving the same electrons that form the fermi surface , or another , independent degree of freedom ( such as a phonon mode that becomes soft at a structural transition ) .
so long as the properties of the qcp are universal , the low - energy behavior does not depend on its microscopic origin . for @xmath80 ,
the system is composed of two decoupled sets of degrees of freedom : free fermions , and pseudospins governed by @xmath81 , which has the form of a @xmath4 transverse field ising model . at zero temperature ,
the pseudospins undergo a second - order quantum phase transition from a paramagnet to an `` antiferromagnet '' that breaks @xmath82 rotational symmetry at @xmath83 .
this transition is in the @xmath5 dimensional classical ising universality class . at @xmath13
, the pseudospin degrees of freedom are gapped in both the nematic and isotropic phases . at finite temperatures ,
a line of second - order classical @xmath4 ising transitions extends from the qcp in the @xmath84 plane . for non - zero @xmath79 ,
the phase diagram is similar , but exhibits quantitative and qualitative modifications .
[ fig : phase_diagram ] shows the phase diagram , obtained by dqmc , for both @xmath16 and @xmath85 .
the @xmath86 transition between the nematic and isotropic phases remains second order , and extrapolates to a new qcp , shifted relative to @xmath87 .
more striking is the change in the slope with which the phase boundary , @xmath88 , approaches the qcp . for @xmath16 ,
the slope diverges at low temperature , consistent with the expectation for the transverse field ising transition , where @xmath89 with @xmath90 and @xmath91 .
in contrast , for @xmath92 , we find that @xmath93 . on the disordered side of the transition
, we define a crossover line by identifying @xmath94 as the value of @xmath95 at which the nematic susceptibility at fixed @xmath8 has fallen to half of its value at @xmath96 .
@xmath97 , the inverse of @xmath98 , also vanishes linearly with @xmath95 upon approaching the qcp , although its slope is steeper than that of @xmath88 .
the linear behavior of @xmath88 for small @xmath99 is also seen for other model parameters . in fig .
[ phasediagram ] we show the phase diagram for two values of the fermion density , controlled by the chemical potential @xmath73 . as the fermion density is reduced , both @xmath100 and the range over which @xmath88 is linear become smaller , indicating that the effect of the coupling between electrons near the fermi surface and the nematic modes becomes weaker . the fact that @xmath88 appears linear at low temperature for both values of fermion density is consistent with this being a _ universal _ property of the metallic qcp .
we have simulated the hamiltonian ( [ eq : h ] ) with system sizes between @xmath101 and @xmath102 , and temperatures between @xmath103 and @xmath104 .
these simulations do not suffer from the minus sign problem : the fermion determinants obtained from integrating out the spin up and spin down fermions are identical and real , so the product is nonnegative .
global updates of the pseudospin space - time configurations are introduced in order to overcome critical slowing down .
some technical details related to the dqmc simulations are discussed in appendix [ app : dqmc ] . here , we shall focus on the results and data analysis . unless stated otherwise , the model parameters used in the simulations were : @xmath9 , @xmath10 , @xmath11 , @xmath105 . ) , while black diamonds and blue squares represent the coupled problem ( @xmath106 ) at @xmath107 and @xmath108 fermions per site , respectively . ] , @xmath109 , as a function of temperature .
panel a ) shows the susceptibility on a linear scale .
the same data are displayed on a log - log scale in panel b ) .
the black line is proportional to @xmath110 . ]
the two - point function for the nematic order parameter field is defined as @xmath111 here , @xmath112 and the nematic order parameter is defined as @xmath113 , with @xmath114 for @xmath115 , @xmath116 for @xmath117 , and @xmath118 otherwise .
the phase diagram discussed above was determined using the nematic susceptibility and standard finite size scaling techniques , described in appendix [ app : finite ] .
we also define a quadrupolar order parameter , @xmath119 in terms of which the quadrupolar correlation function is @xmath120 note that , while microscopically quite different , @xmath121 and @xmath122 have the same symmetries , and both develop anomalous expectation values in the nematic phase . fig . [
fig : chi qcp ] shows @xmath123 , which diverges upon cooling . examining the data on a log - log plot , [ fig .
[ fig : chi qcp](b ) ] , the susceptibility is seen to follow a power law , @xmath124 with @xmath125 , over more than a decade of temperature between @xmath126 .
the slope of @xmath127 vs. @xmath128 is independent of the system size for @xmath129 .
, as a function of @xmath130 for system sizes @xmath131 , and @xmath132 at various temperatures .
( b ) the same data , scaled appropriately for @xmath133 and @xmath134 , collapse on a single curve . the black curve is the function @xmath135 . ]
vs @xmath136 for all orientations of @xmath20 , at @xmath96 and various temperatures .
( b ) shows the same for fixed temperature @xmath137 and a variety of values of the quantum tuning parameter @xmath95 . in each case
the momentum dependence remains quadratic .
momenta from multiple system sizes ( represented by squares for @xmath138 , circles for @xmath64 , and diamonds for @xmath102 ) fall on the same curve , indicating that the properties shown are in the thermodynamic limit . ] .
multiple system sizes are shown with squares for @xmath138 , circles for @xmath64 , and diamonds for @xmath102 .
the error bars reflect uncertainties due to finite system sizes , estimated from the differences between the measured susceptibilities of systems with different boundary conditions . ] at various temperatures .
multiple system sizes are shown with squares for @xmath138 , circles for @xmath64 , and diamonds for @xmath102 .
the error bars reflect uncertainties due to finite system sizes , estimated from the differences between the measured susceptibilities of systems with different boundary conditions . ] for all orientations of @xmath20 at @xmath96 and @xmath8 from @xmath139 to @xmath140 .
momenta from multiple system sizes ( represented by squares for @xmath138 , circles for @xmath64 , and diamonds for @xmath102 ) fall on the same curve , indicating that the properties shown are in the thermodynamic limit .
error bars are smaller than the symbol size . ] to establish the scaling behavior of @xmath141 away from @xmath14 , we plot the susceptibility for different temperatures and system sizes , as a function of @xmath130 [ fig . [
fig : chi h t](a ) ] .
for the system sizes displayed ( @xmath142 ) , @xmath141 is only weakly dependent on @xmath143 . in fig .
[ fig : chi h t](b ) we plot @xmath144 for @xmath145 as a function of @xmath146 , with @xmath147 and @xmath148 .
the different curves collapse onto each other , as expected for a scaling function .
next , we examine the momentum dependence of the static correlator @xmath149 near the qcp .
[ fig : chi h t k ] shows @xmath150 as a function of @xmath151 for various values of @xmath130 and @xmath8 . for small momentum ,
the momentum dependence consists of an essentially isotropic term , approximately proportional to @xmath151 .
combined with the @xmath95 and @xmath8 dependences discussed above , we deduce that the static nematic correlator is well - described by eq .
, with the aforementioned exponents , and with @xmath152 , @xmath153 , @xmath154 , and the lattice spacing set to unity .
this is explicitly demonstrated in fig .
[ fig : static ] .
( in the figure , @xmath155 is the functional approximant to @xmath156 , defined in eq .
[ eq : d_dynamics ] . ) in order to evaluate the consistency of the scaling form , eq .
( [ eq : scaling ] ) , we have also measured the quadrupolar correlator , @xmath38 , over a similar range of parameters as @xmath39 . in fig .
[ fig : comparechibeta ] , we compare the nematic and the quadrupolar susceptibilities at @xmath96 as a function of temperature . like @xmath39 , @xmath38 varies approximately linearly as a function of @xmath110 , although there are large error bars at low temperatures . in fig .
[ fig : comparechih ] we show the @xmath95 dependence of @xmath157 and @xmath158 at various temperatures . in both cases ,
the @xmath95 dependence is approximately linear for small @xmath130 .
however , the range of @xmath130 in which @xmath157 is linear is larger than the corresponding range of linearity of @xmath158 , and the slope of @xmath157 at small @xmath130 is approximately @xmath8 independent , while for @xmath158 it is noticeably temperature dependent .
the momentum dependence of the two correlators is shown in fig .
[ fig : comparechiq ] . both appear isotropic , depending only on @xmath151 .
however , @xmath158 , in contrast to @xmath157 , has noticeable downward curvature .
@xmath157 depends on temperature through an essentially momentum - independent shift , while the temperature dependence of @xmath158 is more complicated . for more details
see appendix e. for a variety of small momenta @xmath20 in an @xmath64 system at temperatures @xmath159 ( shown as squares , stars , circles , and diamonds respectively ) .
different colors represent different values of @xmath20 , and filled symbols mark frequencies @xmath160 where @xmath161 is the minimum value of the bare fermi velocity .
the frequency dependence is essentially linear except at the few smallest frequencies and momenta .
the inset shows the subset of the data with @xmath162 and momenta along the @xmath163 direction shifted by their zero frequency value , with the same scale as the main figure . ] for a variety of small momenta in an @xmath64 system at temperatures @xmath159 ( shown as squares , stars , circles , and diamonds respectively ) .
different colors represent different values of @xmath20 , with the same color scale as in fig .
[ fig : chi w k ] .
filled symbols mark frequencies @xmath160 , where @xmath161 is the minimum value of the bare fermi velocity .
the inset shows the subset of the data with @xmath162 and momenta along the @xmath163 direction shifted by their zero frequency value , with the same scale as the main figure .
the data for @xmath164 are scaled down by a factor of @xmath165 to appear on the same scale . ]
the dependence of @xmath39 on matsubara frequency , @xmath166 , is shown in fig .
[ fig : chi w k ] . at intermediate @xmath21 ( @xmath167 ) and for @xmath168
, @xmath157 is an approximately linear function of @xmath169 with a slope that is independent of @xmath8 , and also independent of both the direction and magnitude of @xmath20 .
however , at the smallest non - zero momenta , a different frequency dependence is visible for @xmath170 .
this is emphasized in the inset of fig .
[ fig : chi w k ] , where @xmath171 has been subtracted .
the frequency dependence at @xmath172 is different from the behavior seen for non - zero @xmath20 ; for @xmath173 , @xmath174 appears to saturate to a frequency independent value , followed by a sudden drop at @xmath175 . the frequency dependence of @xmath157 away from the critical point , shown in fig .
[ fig : chiwk_gapped ] in the appendix [ app : dump ] , is similar to the behavior of @xmath157 at criticality , apart from a constant positive shift ( that corresponds to the fact that we are on the disordered side of the transition , and hence @xmath176 remains finite in the thermodynamic limit ) . the bulk of the data at small values of @xmath130 , @xmath8 , @xmath20 , and @xmath21 is well described by the simple `` functional approximant '' @xmath177 , where @xmath178 as already shown in fig .
[ fig : static ] , this approximant is extremely accurate for all the data at @xmath175 . as can be seen from fig .
[ fig : chi huge ] in appendix e , it does an equally good job for the dynamical response for @xmath20 not too small , but fails to describe the observed dynamics for the few lowest values of @xmath179 .
the equal time correlator gives complementary information about the dynamics , since it integrates over all frequencies .
moreover , `` equal time '' is the same in both real and imaginary time .
thus , in fig . [
fig : chi r ] we exhibit the inverse of the equal time correlator , @xmath180 for @xmath96 at @xmath181 , as a function of distance for various values of @xmath143 .
note that for the largest size system ( @xmath102 ) , @xmath182 , consistent with what one would get by fourier transform of @xmath183 .
@xmath184 is shown for all orientations of @xmath185 , so the narrow spread of points shows the high degree of isotropy exhibited by the data .
note also that for relatively small @xmath186 , the data is independent of @xmath143 , at least for @xmath187 , but that at longer distances the data for the smaller systems deviate systematically from the @xmath188 behavior .
the frequency dependence of @xmath38 is shown in fig .
[ fig : chi_ff ] .
it is clearly qualitatively different from the behavior of @xmath39 in fig .
[ fig : chi w k ] , and correspondingly can not be approximated by any expression of the form in eq .
[ eq : d_dynamics ] .
the uniform quadrupolar density is nearly conserved , as is apparent from the very large values of @xmath189 at nonzero frequencies , shown in the inset .
defined in eq .
( [ eq : kxx ] ) for @xmath190 , @xmath96 , and various values of system size , normalized by the universal value @xmath191 .
momenta from different system sizes lie on the same curve , illustrating the thermodynamic limit . at small values of @xmath192 , @xmath193 lies well below @xmath191 , indicating the system is not superconducting . ]
wave and @xmath194wave channels , @xmath195 [ eq . ( [ eq : sc_suscept ] ) ] , as a function of @xmath95 , at @xmath196 .
different system sizes are displayed .
( c , d ) @xmath195 as a function of temperature at the qcp , @xmath197 . ]
we have probed for superconducting correlations in the vicinity of the qcp by measuring the superfluid density and the superconducting susceptibility , as a function of @xmath95 and @xmath8 .
the superfluid density is given by @xcite @xmath198 where @xmath199 , \label{eq : kxx}\ ] ] and @xmath200 is the current - current correlator @xmath201 here , the current density operator is given by @xmath202 , where @xmath203 .
@xmath204 describes the response of the system to a static orbital magnetic field @xmath205 : @xmath206 where @xmath207 is the vector potential in an appropriate gauge .
extrapolating @xmath208 to the thermodynamic limit is challenging due to the presence of strong finite size effects associated with the fermi surface .
these finite size effects are dramatically reduced by performing the computations in the presence of a weak a uniform magnetic field @xcite .
we adopt this procedure , but make the field spin - dependent , with the total flux through the system being @xmath209 ( @xmath210 ) for spin up ( spin down ) . as shown in ref .
@xcite , such a procedure does not introduce a sign problem for the monte carlo .
moreover , in the limit of large system size , the magnetic field vanishes .
( see appendix [ app : dqmc ] for the implementation of this effective magnetic field in the simulations . )
[ fig : rho_s ] shows @xmath211 as a function of @xmath192 , for different system sizes at temperature @xmath212 and @xmath96 . the data collapse onto a single curve that extrapolates to a value near zero in the limit @xmath213 .
the extrapolated value is clearly much below the universal value at the bkt transition , @xmath214 .
this implies that at @xmath215 , the system is not superconducting down to @xmath216 .
similarly , the superfluid density for @xmath217 is found to be below @xmath218 for all @xmath219 .
note that , from eq .
( [ eq : kxx ] ) , @xmath220 is the orbital susceptibility . from fig .
[ fig : rho_s ] , we see that in our system , this quantity is negative , hence the orbital response is paramagnetic @xcite .
this indicates that there are no strong superconducting fluctuations ( which would produce diamagnetism ) .
we have also computed the superconducting susceptibilities in the @xmath221wave ( @xmath222 ) and @xmath194wave ( b@xmath223 ) channels , as a function of @xmath95 and @xmath8 : @xmath224 here , @xmath225 and @xmath226 , where @xmath227 is defined below eq .
( [ eq : chi1 ] ) .
the superconducting susceptibilities are displayed in fig .
[ fig : p_of_t ] . as expected for a non - superconducting state , both @xmath228 and @xmath229 saturate as a function of system size at fixed @xmath8 and @xmath95
[ see fig .
[ fig : p_of_t](a , b ) ] . @xmath228 has a maximum at @xmath230 , while @xmath229 is smaller and more or less independent of @xmath95 .
the enhancement of @xmath228 by nematic fluctuations is further reflected in its behavior as a function of temperature at a fixed @xmath95 [ fig .
[ fig : p_of_t](c , d ) ] .
@xmath228 rises more rapidly than logarithmically at low temperatures .
in contrast for non - interacting electrons @xmath231 , a functional form consistent with the @xmath8 dependence of @xmath229 at low @xmath8 . we interpret this as an indication that the @xmath221wave susceptibility is enhanced by induced attractive interactions mediated by nematic fluctuations ; the ground state for @xmath95 near @xmath14 is likely an @xmath221wave superconductor , although with @xmath232 .
having found that the system is not superconducting down to the temperatures reached in this study , we examine the single - fermion green s function to glean information about the metallic state in the vicinity of the qcp .
the metal is naturally characterized via the single particle density of states @xmath233 , the fermi velocity @xmath234 , and the quasiparticle weight @xmath235 . however
, measuring these quantities within dqmc is not straightforward , since they generally require real - frequency information inaccessible without analytic continuation .
in addition , @xmath234 and @xmath235 are well defined only at zero temperature , while our simulations are at finite temperature .
we partially resolve these difficulties by defining finite temperature objects @xmath236 , @xmath237 , and @xmath238 which can be measured numerically , and whose zero temperature limits are @xmath239 , @xmath234 , and @xmath235 respectively . readers interested in the raw frequency dependence of the fermionic green s function and self - energy
are referred to appendix [ app : g_and_sigma ] .
real frequency information can be extracted from our imaginary - time data using the identity @xcite : @xmath240 where @xmath241 is the fermion spectral function .
@xmath242 is evidently equal to the spectral function integrated over a range of energies @xmath243 .
we therefore define @xmath244 @xmath245 |_{{\bf k}={\bf k_f},\tau=\beta/2 } , \label{vf}\end{aligned}\ ] ] @xmath246 we explain in appendix [ app : metallic ] how these quantities attain the desired zero temperature limits .
we also discuss in the appendix the procedures used to estimate these quantities in finite size systems with discrete momentum and imaginary time .
[ fig : dos ] shows @xmath247 for an @xmath64 system as a function of @xmath95 , for different temperatures .
@xmath247 remains nonzero for all values of @xmath95 , consistent with a finite density of states in the limit @xmath248 .
this indicates the absence of a substantial gap in the fermionic spectrum , even a gap with nodal points on the fermi surface , down to @xmath249 , corroborating the conclusion of sec .
[ sec : sc ] that the system is not superconducting down to these temperatures . moreover , the fact that @xmath247 seems to neither vanish nor diverge at low temperature and @xmath215 has important implications for the nature of the metallic state of the qcp . were the fermions to acquire a positive ( negative ) anomalous dimension , @xmath250 , the density of states would go to zero ( diverge ) at @xmath13 @xcite .
we elaborate on this point in sec .
[ sec : discussion ] . instead , the saturation of the density of states is consistent with a fermi liquid , at least down to the temperature scales in our simulations .
note , as well , that the density of states appears to be depressed with increasing nematic order for @xmath251 , defined in eq .
( [ eq : ntilde ] ) , as a function of @xmath95 for system size @xmath64 at different temperatures .
@xmath252 equals the density of states at the fermi energy in the limit @xmath248 . ]
[ defined in eq .
[ eq : zdef ] ] as a function of angle @xmath253 along the fermi surface , for different values of @xmath95 .
the qcp is at @xmath254 .
right panels show the temperature dependence of @xmath238 for high symmetry directions of @xmath255 .
@xmath256 approaches the conventionally defined quasiparticle weight as @xmath44 . ] as a function of @xmath95 for different temperatures .
circles ( diamonds ) correspond to @xmath257 ( @xmath258 ) , respectively ( where @xmath253 is defined in the inset of fig .
[ fig : z ] ) .
( b ) the fermi velocity renornalization , @xmath259 ( where @xmath161 is defined in eq .
[ vf ] and @xmath260 is the non - interacting fermi velocity ) as a function of @xmath95 for @xmath258 , @xmath261 . ] the results for @xmath262 , where @xmath253 is the angle between @xmath263 and the @xmath264 axis , are shown in fig .
[ fig : z ] for different values of @xmath265 . at @xmath266 ,
@xmath256 is close to unity at all temperatures .
this can be understood as a consequence of the fact that at @xmath266 , the coupling between fermions and the nematic critical modes vanishes by symmetry ; hence , at these `` cold spots '' , the effects of scattering off the critical modes are minimal .
@xmath256 decreases systematically as @xmath253 approaches @xmath267 , and as @xmath95 approaches @xmath14 .
@xmath268 decreases with decreasing temperature with negative curvature . on the basis of the present data , it is difficult to say whether it extrapolates to a finite value in the @xmath248 limit ( indicating a fl ground state ) , or vanishes with a small exponent , @xmath269 where @xmath270 ( indicating a breakdown of fl behavior )
. it would be difficult to reconcile the data with @xmath271 larger than about @xmath272 .
[ fig : vf](a ) shows the same data as a function of @xmath265 for @xmath258 and @xmath273 , emphasizing the decrease in @xmath256 as @xmath95 approaches @xmath14 .
measuring @xmath256 close to the qcp for @xmath251 ( in the nematic phase ) has proven difficult , since the two configurations of the order parameter correspond to two different fermi surfaces with a small splitting between them . in fig .
[ fig : vf](b ) we show the renomalization of the fermi velocity , @xmath259 , where @xmath260 is the fermi velocity for the non - interacting band structure ( @xmath80 ) .
as with @xmath256 , @xmath161 is hardly renormalized at the cold spot @xmath257 .
in contrast , at @xmath258 , @xmath161 is strongly reduced from its non - interacting value , and the velocity renormalization increases as @xmath95 approaches @xmath14 . however , down to @xmath274 , there is no obvious indication that @xmath161 vanishes ( i.e. , the effective mass diverges ) as @xmath275 , even at @xmath215 . over the temperature range explored in this study , the single fermion green s function
is consistent with a renormalized fermi liquid , with an angle- and @xmath95- dependent quasi - particle weight and velocity .
however , the expected differences between a fermi liquid and a `` marginal fermi liquid '' or a weak ( small @xmath271 ) non fermi liquid are subtle at non - zero @xmath8 .
we can rule out a non - fermi liquid with a substantial value of @xmath271 and/or @xmath250 , but not one with @xmath276 and @xmath250 sufficiently small .
we view our results as a numerical `` experiment . '' for an explicit lattice hamiltonian , we have obtained detailed and extensive `` measurements '' of physical quantities . in the neighborhood of the quantum phase transition we find an enhanced tendency to superconductivity , substantial scattering of the fermionic quasiparticles , and a clearly delineated quantum critical fan .
moreover , for a large range of parameters , a simple scaling form provides a good description of the thermodynamic nematic fluctuations . in laboratory experiments on solid state systems , good scaling collapse over a similar dynamical range
is often interpreted in terms of universal critical exponents . however , in our system , the fermion bilinear quadrupolar correlations are only consistent with the apparent scaling form of the nematic correlator over a narrow region as a function of @xmath277 .
this is reason to question whether the behavior we are seeing is characteristic of the asymptotic approach to a quantum critical point .
it is possible that the success of the scaling analysis is fortuitous .
alternatively , the observed scaling may reflect an intermediate asymptotic regime , possibly associated with an unstable multicritical point .
it is also possible that corrections to universal scaling behavior are simply smaller for @xmath39 than for @xmath38 , in which case the scaling form of @xmath39 might reflect the true critical behavior . at the very least ,
the observation of simple power laws invites a comparison with theories of metallic quantum criticality .
the theory of metallic qcps , along with the closely related theory of metals coupled to fluctuating gauge fields@xcite , has traditionally been treated by integrating out the fermions to obtain an effective action for the bosonic modes that is highly non - local in both space and time . in this `` hertz - millis '' approach , a renormalization group ( rg ) analysis
is then undertaken for the effective field theory so defined .
while much can be learned from this program , there are at least two general causes for concern : 1 ) the analyticity of the @xmath278-function , which is a core ingredient of rg , is typically ensured by the locality of the action . for a non - local action ,
all aspects of the analysis , especially the enumeration of potentially relevant interactions that can be generated under rg , need to be tested explicitly .
2 ) in many cases , including in particular the 2d nematic qcp @xcite , if one considers the back effect of the collective fluctuations on the fermions , one finds that fermi liquid theory is perturbatively unstable .
the question then arises concerning the self - consistency of the effective action which is obtained by integrating out an assumed fermi liquid .
nonetheless , there is a general feeling that this approach is reliable in large enough spatial dimension @xcite , as well , the kondo problem can be mapped@xcite to the problem of a two - level system coupled to an ohmic heat bath , for which the non - local effective action is equivalent to that of a @xmath279 classical inverse - square ising model .
this problem can , in turn , be treated using a suitable rg analysis @xcite . ] and its main results are similar in structure to those of many subsequent studies .
accordingly , we briefly review the main features of the hertz - millis treatment of the ising nematic qcp . in this context
, the quadratic part of the effective action for the nematic mode is of the form @xmath280 where @xmath281 and @xmath24 are ( renormalized ) functions of @xmath8 and @xmath95 , and @xmath282 is the suitably defined particle - hole response function ( with the zero - frequency piece subtracted ) @xmath283 } { [ \varepsilon_{\mathbf{k}+\mathbf{q}}-\varepsilon_{\mathbf{k}}][i\omega_n+\varepsilon_{\mathbf{k}+\mathbf{q}}-\varepsilon_{\mathbf{k } } ] } , \label{q}\ ] ] where @xmath284 is a dimensionless form factor of appropriate symmetry , whose form depends on microscopic details ( e.g. @xmath285 ) .
the essential features of this expression relate to its asymptotic ( ir ) behavior in the limit @xmath286 , @xmath287 where @xmath288 is the point on the fermi surface ( assumed to be unique mod inversion ) at which the fermi velocity , @xmath289 , is perpendicular to @xmath290 .
( the fact that the integral in eq .
( [ q ] ) for given @xmath20 is dominated by the neighborhood of this point on the fermi surface is the essential observation motivating the `` patch '' theories of this problem . )
the hertz - millis approach implies that the dynamical exponent @xmath291 at tree level . from this , a naive scaling analysis ( neglecting the possibility of higher - order non - local interactions ) concludes that , since @xmath292 is well above the ising upper critical dimension , mean - field exponents @xmath293 , @xmath134 , and @xmath294 apply , consistent with what we find .
( see table i. ) moreover , one would expect violations of scaling ( for instance @xmath295 and @xmath296 ) . indeed in refs . @xcite it is shown that at criticality , such an analysis leads to @xmath297 $ ] .
as we do not have the dynamical range to detect a logarithm , this , too , is consistent with our results .
however , on the face of it , our results for the dynamical nematic response are inconsistent with a dynamical exponent @xmath291 and with the sort of anisotropies one might expect from the patch construction .
moreover , the relative weakness of any deviations of the single - particle properties from fermi liquid behavior appears inconsistent with the hertz - millis predictions defined here has not been explicitly computed within the hertz - millis approach . applying @xmath298 scaling to the @xmath13 self energy ( which , of course may not be valid ) yields @xmath299 , a noticeably more rapid variation than the data of fig .
[ fig : z]d . ] . on the other hand , at the smallest momenta , deviations of @xmath39 from the scaling form , @xmath300 , eq .
( [ eq : d_dynamics ] ) , become apparent .
this may indicate a crossover connecting fermi surface copies in adjacent brillouin zones ( see fig .
[ fig : fermi_surface ] ) , and this might lead to a crossover scale at small momentum . ] from an intermediate @xmath301 regime to a different behavior in the deep infra - red @xcite .
the foregoing discussion of the relation between our data and hertz - millis theory can be applied to numerous subsequent attempts to bring the problem under theoretical control . among the approaches employed by these studies , whose results have similar structure to those of hertz - millis , are large @xmath302 methods ( where @xmath302 is the number of fermion flavors)@xcite , dimensional@xcite and dynamical@xcite regularization , and higher dimensional bosonization@xcite .
the related problem of orbital loop current criticality in a metal has been argued@xcite to lead to a critical bosonic correlator with @xmath303 frequency dependence ( as seen in our results ) , but with an asymptotic absence of momentum dependence ( not seen in our results ) characteristic of local quantum criticality " @xcite . in the vicinity of @xmath304 , several varieties of intermediate asymptotics corresponding to different schemes of large @xmath302 and @xmath305 ( the number of boson flavors ) have also been explored@xcite , at least one of which has @xmath306 upon naive extrapolation to @xmath4 .
we believe our results motivate revisiting existing theories of nematic quantum criticality in metals . in particular , the isotropy of bosonic correlations we observe appears to be at odds with patch " constructions .
finally , there has been considerable theoretical interest in the issue of superconductivity in the neighborhood of a nematic quantum critical point .
it has been argued @xcite that superconductivity with a critical temperature of order 1 times microscopic scales should be expected near the nematic quantum critical point .
indeed , it has been suggested that the superconducting @xmath307 at criticality may be so high that it preempts the quantum critical regime in which non - fermi liquid behavior would be expected .
more modestly , in the small @xmath79 limit , it has been shown@xcite that nematic fluctuations do indeed mediate attractive pairing interactions , and that the resulting transition temperature grows singularly as the quantum critical point is approached either from the ordered or the disordered side . in agreement with this latter result , the superconducting susceptibilities shown in fig .
[ fig : p_of_t ] do show a maximum at @xmath96 .
however , if @xmath307 at criticality is a number of order 1 times the fermi energy , that number is apparently sufficiently small that @xmath308 for our chosen parameters . at larger values of the boson - fermion coupling constant ,
preliminary results @xcite show that the ground state is an @xmath309wave superconductor whose @xmath307 is maximal near the nematic qcp .
there is increasingly extensive evidence that electron nematic phases are common in highly correlated electronic fluids.@xcite in particular , it has been suggested that a nematic qcp occurs near optimal doping in both the cuprates@xcite and in the fe - based superconductors@xcite .
however , the experimental situation in all these cases is complicated , and in some cases controversial .
it is premature to attempt any sort of serious comparison between the present results and the experiments in these materials .
however , the experimental case is clearest in certain fe based superconductors , so it is worth noting a few points of comparison .
most strikingly , elastoresistance @xcite , raman scattering @xcite , and elastic constant @xcite measurements show a large range of @xmath8 and doping over which a large nematic susceptibility can be documented with a remarkably systematic currie weiss @xmath8 dependence @xmath310 $ ] where @xmath311 appears to depend roughly linearly on `` doping '' concentration , @xmath264 , and to pass through zero at a critical doping concentration , @xmath312 , which approximately coincides with the optimal doping for superconductivity .
it is impossible not to be encouraged by the similarity between this experimental finding and our numerical results .
note added : after our paper was posted , a new theoretical analysis of the finite @xmath8 dynamics near a nematic metallic qcp @xcite has found a possible route to understanding the observed behavior of @xmath39 , and in particular for the apparent dynamical exponent @xmath306 .
discretizing imaginary time as @xmath313 , the grand partition function becomes @xmath314 where @xmath315 here , @xmath316 is a row vector of spin @xmath65 fermionic creation operators for the spatial sites @xmath317 , where @xmath143 is the linear dimension . @xmath318 and @xmath319 are the terms in eq .
[ eq : hfhb ] proportional to @xmath320 and @xmath73 respectively .
we use a checkerboard decomposition to describe the kinetic energy matrices @xmath321 , whose elements are @xmath322 , i.e , @xmath323 enumerates the horizontal / vertical bonds originating from a site with an even / odd index .
we allow @xmath324 to depend on spin in order to implement the spin - dependent magnetic field discussed below . plugging in unity operators in the @xmath325 sector at every time slice and taking the trace over the fermions , we bring the partition function to a form which can be sampled using monte - carlo techniques : @xmath326 where @xmath327 is a matrix which depends implicitly on the c - numbers @xmath328 through the relation @xmath329 , and the bosonic part of the action
is given by @xmath330 we have used @xmath331 in our simulations .
monte - carlo sampling is most efficient when the determinant in is positive
. here , the absence of a sign problem is be guaranteed by microscopic time reversal symmetry . to see this , note that the matrices @xmath332
are block diagonal in the spin sector , and so the total determinant is the product of the spin up and spin down determinants .
if time reversal is preserved , i.e
@xmath333 then the product of the determinants is clearly positive .
it has long been known@xcite that finite size effects in fermionic systems , in particular at low temperatures , can be greatly improved by the application of a weak orbital magnetic field . on a finite size system with periodic boundary conditions
, the field must be such that an integer multiple of the flux quantum passes through the system , @xmath334 . in order to preserve time reversal symmetry as required in , in this work
we have applied the opposite field for both spin species , such that @xmath335 .
such a field vanishes in the thermodynamic limit .
although our simulations do not suffer from the sign problem , close to the quantum critical point our simulations do suffer from critical slowing down . to avert this
, we found it was helpful to include global monte - carlo updates , in addition to local ( metropolis ) ones .
our global updates consisted of a slight modification to the wolff algorithm@xcite : starting from a space - time configuration of pseudo - spins , a cluster of pseudo - spins is constructed using the usual wolff algorithm , supposing our action is given by , i.e without taking into account the coupling to the fermions .
next , we propose flipping the pseudo - spins which belong to the cluster , and accept the move with a probability which is the ratio of the determinants in between the two configurations .
it can easily be seen that such a move obeys detailed balance .
the finite temperature phase boundary @xmath88 ( or equivalently @xmath336 ) can be computed by using _
classical _ finite size scaling techniques for two dimensional ising transitions , characterized by susceptibility exponent @xmath337 and and correlation length exponent @xmath338 . at a given temperature and for a given system size @xmath143 ,
the most singular part of the thermodynamic susceptibility , @xmath339 , satisfies @xmath340 @xmath341 can therefore be identified using a single parameter data collapse , as illustrated in fig .
[ fig : classical_collapse ] . at low temperatures , the region of parameter space corresponding to classical criticality narrows , so that larger system sizes are needed to reliably estimate @xmath341 . for a given temperature . here , for @xmath342
, the best data collapse is obtained for @xmath343 .
the exponents take the two dimensional ising values @xmath337,@xmath338 for this thermal transition .
the nematic susceptibility and quadrupolar susceptibility have similar behavior.,title="fig : " ] for a given temperature . here , for @xmath342 , the best data collapse is obtained for @xmath343 .
the exponents take the two dimensional ising values @xmath337,@xmath338 for this thermal transition .
the nematic susceptibility and quadrupolar susceptibility have similar behavior.,title="fig : " ] for @xmath59 .
data points from different temperatures mostly lie on the same curve .
right panels : the imaginary part of the self energy .
the dashed line in the bottom left panel corresponds to @xmath344 . ]
write the imaginary time green s function in a spectral representation : @xmath345 here , @xmath346 is a many - body eigenstate with energy @xmath347 , and @xmath348 is the grand partition function . in this representation ,
the fermion spectral function is @xmath349 we can write eq .
( [ eq : g1 ] ) as @xcite @xmath350 if the ground state is a fermi liquid , we can extract the quasi - particle weight and fermi velocity renormalization from eq .
( [ eq : g3 ] ) .
we assume that at low temperatures , the spectral function in the vicinity of the fermi surface obtains a fermi liquid form , @xmath351 , where @xmath352 is the quasi - particle contribution and @xmath353 is a regular background . in the limit @xmath248 , @xmath352 becomes a delta - function like peak : @xmath354 , where @xmath355 is the quasi - particle dispersion near the fermi surface , @xmath256 is the quasi - particle weight , @xmath356 is the fermi velocity . at finite but small @xmath8 and @xmath357 , the quasi - particle peak at @xmath263 obtains a width that scales as @xmath358 .
the regular part satisfies @xmath359 $ ] . inserting the fermi liquid form of @xmath62 into eq .
( [ eq : g3 ] ) , we obtain that at low temperatures , @xmath360 we can use this relation to estimate @xmath256 and @xmath361 along the fermi surface . for a point on the fermi surface , @xmath362 .
the dispersion can be estimated from @xmath363 / d \tau |_{\tau = \beta/2}$ ] , and the fermi velocity is then evaluated according to @xmath364 . some caution is needed because in our finite size systems , the values of @xmath57 are quantized , and the @xmath365 grid does not in general intersect the fermi surface .
one can instead estimate @xmath366 from the points nearest to the fermi surface .
also , our measurements of @xmath367 are taken at finite temperatures , yielding the finite temperature estimators @xmath247 , @xmath368 and @xmath369 discussed in the text ( equations [ eq : ntilde ] , [ vf ] , and [ eq : zdef ] ) .
an extrapolation to the limit @xmath370 is necessary in order to deduce the
true " fermi liquid properties , to the extent they are well defined .
in fig . [ fig : matsubara_g ] we show the matsubara frequency dependence of the inverse fermion green s function .
linear frequency dependence , consistent with a fermi liquid , seems to dominate down to the lowest frequencies available .
finer details are visible by considering the self energy ( right panels ) which shows both a noticeable temperature dependence and a deviation from linearity for @xmath371 .
[ fig : comparechiq ] of the main text shows the momentum dependence of @xmath157 and @xmath158 at @xmath45 and temperatures @xmath372 . the momentum dependence at low temperatures is shown in fig .
[ fig : comparechiq2 ] , here plotted versus @xmath179 instead of @xmath151 . whereas @xmath157 is quadratic in @xmath179
, @xmath158 is apparently linear in @xmath179 at low temperatures over a substantial range of momenta , a finding we have not yet understood . figs . [
fig : chi w k ] and [ fig : chi_ff ] of the main text show @xmath157 and @xmath158 versus frequency for a variety of momenta , at low temperature and @xmath45 .
[ fig : chiwk_gapped ] shows similar data for @xmath373 , illustrating that deviation from criticality appears mostly through an additive shift of the inverse correlator .
[ fig : static ] of the main text shows the validity of the functional approximant @xmath374 of eq . in describing thermodynamic
nematic correlations @xmath375 . fig .
[ fig : chi huge ] plots @xmath157 vs @xmath374 for all frequencies .
the agreement is very good except for the smallest momenta , shown in blue and black .
[ fig : chi r ] of the main text shows the position dependence of the equal time nematic correlator @xmath376 . the analogous data for the equal time quadrupolar correlator @xmath377 are somewhat more messy .
this can be seen in fig .
[ fig : chiq r ] where we plot the dependence of @xmath378 for @xmath185 along the symmetry directions , @xmath379 and @xmath380 ; as is apparent , the error bars ( associated with the sensitivity of the results to boundary conditions ) are sufficiently large that it is difficult to conclude much from this figure .
( top ) and inverse quadrupolar correlator @xmath381 ( bottom ) at @xmath382 for a variety of small momenta @xmath20 in an @xmath64 system at temperatures @xmath159 ( shown as squares , stars , circles , and diamonds respectively ) . here
, there is a non - vanishing energy scale associated with nematic fluctuations since we are deep in the disordered phase , but the frequency dependence is similar to that at @xmath96 ( see fig .
[ fig : chi w k ] ) .
different colors represent different values of @xmath20 , and filled symbols mark frequencies @xmath160 .
the insets show the subset of the data with @xmath162 and momenta along the @xmath163 direction shifted by their zero frequency values . for the inset of the bottom panel
the data at @xmath164 are scaled down by a factor of @xmath383 .
, title="fig : " ] ( top ) and inverse quadrupolar correlator @xmath381 ( bottom ) at @xmath382 for a variety of small momenta @xmath20 in an @xmath64 system at temperatures @xmath159 ( shown as squares , stars , circles , and diamonds respectively ) . here
, there is a non - vanishing energy scale associated with nematic fluctuations since we are deep in the disordered phase , but the frequency dependence is similar to that at @xmath96 ( see fig .
[ fig : chi w k ] ) .
different colors represent different values of @xmath20 , and filled symbols mark frequencies @xmath160 .
the insets show the subset of the data with @xmath162 and momenta along the @xmath163 direction shifted by their zero frequency values . for the inset of the bottom panel
the data at @xmath164 are scaled down by a factor of @xmath383 .
, title="fig : " ] | the ising nematic quantum critical point ( qcp ) associated with the zero temperature transition from a symmetric to a nematic _ metal _ is an exemplar of metallic quantum criticality .
we have carried out a minus sign - free quantum monte carlo study of this qcp for a two dimensional lattice model with sizes up to @xmath0 sites .
for the parameters in this study , some ( but not all ) correlation functions exhibit scaling behavior over the accessible ranges of temperature , ( imaginary ) time , and distance , and the system remains non - superconducting down to the lowest accessible temperatures . the observed scaling behavior has remarkable similarities with recently measured properties of the fe - based superconductors proximate to their putative nematic qcp . |
capacity characterization of multi - source networks is one of the fundamental problems in network information theory .
however the capacity is not fully characterized even for the simplest setting of the two - user interference channel @xcite that leads to surging interests and demands on approximate capacity characterization .
recently there has been a series of significant progress on approximate capacity characterization @xcite .
the capacity region of the two - user gaussian interference channel was characterized by etkin , tse , and wang within one bit precision @xcite and the sum capacity @xmath5 of the @xmath0-user time - varying gaussian interference channel was characterized as @xmath6 by cadambe and jafar @xcite , where @xmath7 denotes the signal to noise ratio ( snr ) .
that is , the degrees of freedom ( dof ) or capacity pre - log term of the @xmath0-user interference channel is given by @xmath8 . to achieve @xmath8 dof ,
a new interference management technique called the interference alignment was used , which minimizes the dimensions occupied by interference at destinations by aligning the interference from multiple unintended sources .
the interference alignment can also be used to compute achievable dof of some other channels such as the @xmath0-user multiple - input multiple - output ( mimo ) channel @xcite , the deterministic @xmath0-user interference channel @xcite , and the @xmath9-network @xcite in which each of @xmath10 sources has messages for @xmath11 destinations , i.e. , total of @xmath12 message sets . in this paper , we consider relay networks consisting of multiple sources , multiple corresponding destinations , and multiple relays .
relays have been traditionally used for extending coverage in wireless environments , e.g. , amplify - and - forward ( af ) based relays .
although the dof is upper bounded by @xmath13 for fully connected @xmath0-user interference channels @xcite , with help of relays it may be possible to improve the dof also .
the work @xcite has applied their @xmath9 network results to a two - hop network with @xmath10 sources @xmath11 destinations with @xmath2 relays between them . assuming @xmath14 , they showed that @xmath15 dof is achievable .
notice that whereas a trivial upper bound assuming perfect cooperation between relays is @xmath0 if @xmath16 , the achievable dof of @xmath15 converges to @xmath0 only if @xmath17 .
hence one of the basic questions about relay networks is the minimum number of required relays for achieving the optimal @xmath0 dof .
the works @xcite have addressed similar questions and shown that with @xmath0 fixed the sum rate of @xmath18 is achievable if @xmath17 @xcite .
the main contributions of this paper are the follows . * for @xmath0-@xmath2-@xmath0 networks ,
i.e. , @xmath3-hop network with @xmath2 relays , we show that @xmath19 dof is achievable .
hence the optimal @xmath0 dof is achievable if @xmath20 . to show the achievability , interference alignment combined with distributed interference cancellation using multiple relays
is used over multiple symbols ( symbol extension ) to utilize more diversity provided by time - varying channels . a similar interference cancellation technique
called interference neutralization was used for deterministic and non - fading gaussian @xmath3-user @xmath3-hop networks @xcite , where multiple relays are cooperatively used to cancel interference . in our case , such distributed interference cancellation is combined with symbol extension in a more general network . * for @xmath0-hop networks with @xmath0 relays in each layer ,
we show that the optimal @xmath0 dof is achievable if @xmath0 is even and the probability of channel matrix is a function of its frobenius norm only .
we apply a new technique called opportunistic interference cancellation where the relays in each layer delay - amplify - and - forward their received signal vector by waiting for an appropriate channel instance in the next hop such that the overall channel matrix from the sources to destinations become a scaled identity matrix .
hence @xmath0 source destination ( s d ) pairs can communicate concurrently without interference .
this is related to the opportunistic interference cancellation for finite - field networks @xcite , but our scheme for gaussian relay networks in this paper works differently . for both cases , opportunistic interference cancellation is applied and the optimal dof of gaussian networks and the optimal sum rate of finite - field networks are achieved in certain cases .
notice that @xmath21 relays are again required to obtain @xmath0 dof .
the concept of opportunistic channel pairing can be found in @xcite for the finite - field interference channel and in @xcite for the gaussian interference channel , which it was called ergodic interference alignment .
for the multi - hop case , opportunistic interference cancellation was applied for finite - field networks in @xcite .
throughout the paper , we use notations @xmath22 and @xmath23 to denote a matrix and a vector , respectively .
the transpose , conjugate transpose , and frobenius norm of @xmath22 ( or @xmath23 ) are denoted by @xmath24 , @xmath25 , and @xmath26 ( or @xmath27 , @xmath28 , and @xmath29 ) , respectively .
the diagonal matrix having @xmath30 as the @xmath31-th diagonal element , the @xmath32 identity matrix , and the @xmath33 all - zero matrix are denoted by @xmath34 , @xmath35 , and @xmath36 , respectively .
we also use @xmath37 and @xmath38 to denote @xmath39-symbol - extended matrix and vector consisting of @xmath40 $ ] through @xmath41 $ ] , respectively .
that is , @xmath42,\cdots , a[n])$ ] and @xmath43,\cdots , a[n]]^t$ ] .
we study a @xmath1-hop relay network having @xmath44 layers with @xmath45 nodes in the @xmath46-th layer .
the nodes in the first and the last layer are the sources and the destinations , respectively .
thus @xmath47 is the number of s d pairs . for simplicity ,
let us denote the @xmath31-th node in the @xmath46-th layer by node @xmath48 .
for @xmath49 , let @xmath50 $ ] denote the transmit signal of node @xmath48 at time @xmath51 .
then the received signal @xmath52 $ ] of node @xmath53 at time @xmath51 is given by in @xmath52 $ ] means the @xmath46-th hop . ]
@xmath54=\sum_{i=1}^{k_m}h_{ji , m}[t]x_{i , m}[t]+z_{j , m}[t],\ ] ] where @xmath55 $ ] is the channel from node @xmath48 to node @xmath53 at time @xmath51 and @xmath56 $ ] is the noise at node @xmath53 at time @xmath51 .
the noise terms @xmath56 $ ] s are independent and identically distributed ( i.i.d . ) complex gaussian with zero - mean and unit - variance .
we assume time - varying channels such that @xmath55 $ ] s are i.i.d .
drawn from a continuous distribution and @xmath57=h)=0 $ ] for all @xmath58 .
we assume every source and relay has the same power constraint @xmath7 .
let us denote the transmitted and received signal vectors of the @xmath46-th hop by @xmath59=[x_{1,m}[t ] , \cdots , x_{k_m , m}[t]]^t$ ] and @xmath60=[y_{1,m}[t ] , \cdots , y_{k_{m+1},m}[t]]^t$ ] , respectively .
then the input output relation of the @xmath46-th hop can be represented as @xmath61=\mathbf{h}_m[t]\mathbf{x}_m[t]+\mathbf{z}_m[t ] , \label{eq : input_output_vec}\ ] ] where @xmath62 $ ] is the channel matrix of the @xmath46-th hop whose @xmath63-th element is given by @xmath55 $ ] and @xmath64=[z_{1,m}[t],\cdots , z_{k_{m+1},m}[t]]^t$ ] is the noise vector of the @xmath46-th hop .
the channel state information ( csi ) is assumed to be available at all nodes , i.e. , each node knows @xmath65 $ ] to @xmath66 $ ] at time @xmath51 . for simplicity , we omit time index @xmath51 in the rest of the paper . the @xmath31-th source sends a message @xmath67 to its destination at a rate of @xmath68 during @xmath69 channel uses .
the rate tuple @xmath70 is said to be achievable if the probability of error for all s d pairs can be made arbitrarily small by choosing large enough @xmath69 .
the capacity region @xmath71 is the convex hull of the closure of all achievable rate tuples and the sum capacity @xmath5 is the supremum of all achievable sum rates .
then the dof is defined as @xmath72
in this section , we consider a two - hop network with @xmath73 and @xmath74 and assume @xmath16 , which is denoted by the @xmath0-@xmath2-@xmath0 network .
we assume af relaying . because multiple replicas of a transmit signal interfere with unintended destinations through multiple relays
, we can make the replicas cancel each other by appropriately choosing the gains at relays .
since we assume symbol extension , i.e. , we send messages using multiple channel instances , we have vectorized af , i.e. , each relay can multiply its received vector by a matrix .
consider @xmath39 symbol extension .
then the received signal vector of the @xmath75-th relay can be represented as @xmath76 where @xmath77 and @xmath78,\cdots , x_{i,1}[n]]^t\nonumber\\ \bar{\mathbf{y}}_{j,1}\!\!\!\!\!\!\!\!\!\!\!&&=[y_{j,1}[1],\cdots , y_{j,1}[n]]^t\nonumber\\ \bar{\mathbf{z}}_{j,1}\!\!\!\!\!\!\!\!\!\!\!&&=[z_{j,1}[1],\cdots , z_{j,1}[n]]^t\nonumber\\ \bar{\mathbf{h}}_{ji,1}\!\!\!\!\!\!\!\!\!\!\!&&=\diag(h_{ji,1}[1],\cdots , h_{ji,1}[n]).\end{aligned}\ ] ] each relay transmits @xmath39 linear combinations of its @xmath39 received signals to the destinations . specifically ,
the transmit signal vector of the @xmath75-th relay is given by @xmath79 where @xmath80 & \cdots & \gamma_j[1,n ] \\
\vdots & \ddots & \vdots \\
\gamma_j[n,1 ] & \cdots & \gamma_j[n , n ] \\ \end{array } \right]\ ] ] represents the gain matrix of the @xmath75-th relay .
symbols , applies a linear transform , and sends it in the next @xmath39 time slots .
therefore , there will be one block delay at the relays . to simplify notations , we omit block indices . ] similarly
, the received signal vector of the @xmath81-th destination can be represented as @xmath82 where @xmath83 . combining ( [ eq : input_output_1hop ] ) through ( [ eq : input_output_2hop ] ) ,
we obtain @xmath84 notice that the first term is the intended signals and the second term is the interfering signals and the third term is the noise propagation due to af based relaying .
note that the following condition guarantees that the interference from the @xmath31-th source to the @xmath81-th destination ( @xmath85 ) will be nullified .
@xmath86 fig .
[ fig : int_mit ] illustrates this .
the detailed analysis will be given in lemma [ lemma : n_cancel ] .
for transmission , we only use the time slots where channel gains satisfy @xmath87|\leq g_{\max}$ ] for all @xmath31 and @xmath75 , where @xmath88 , @xmath89 , and @xmath90 .
since we assume @xmath91=h)=0 $ ] , the probability of slot utilization can be arbitrarily close to one by setting @xmath92 and @xmath93 as arbitrarily small and large , respectively . as a result introducing @xmath92 and @xmath93
does not affect the dof . for @xmath39 symbol extension ,
we allocate @xmath94 symbols to the first s d pair and @xmath95 symbols to each of the remaining @xmath96 s d pairs .
the first source transmits @xmath97 symbols without transmit beamforming and transmits @xmath98 symbols via beamforming vectors @xmath99 to @xmath100 .
similarly , for @xmath101 , the @xmath31-th source transmits @xmath97 symbols without beamforming and @xmath102 symbols via @xmath103 to @xmath104 .
the interference caused by the @xmath97 symbols will be cancelled by using relay coefficients and the interference caused by the remaining symbols will be aligned by using transmit beamforming .
the following lemma shows that each source can transmit @xmath97 symbols without interfering with unintended destinations if we set @xmath105 .
[ lemma : n_cancel ] suppose a @xmath0-@xmath2-@xmath0 network with @xmath39 time extension .
then there exist @xmath106 s such that each source transmits up to @xmath107 symbols without interfering with unintended destinations .
we refer readers to the full paper @xcite .
then we apply the interference alignment technique to the remaining interference and obtain the achievable dof in the following theorem .
[ th : k_l_k ] suppose a @xmath0-@xmath2-@xmath0 network
. then @xmath108 .
first of all , we briefly discuss the power constraint issue . since each channel used for transmission satisfies @xmath109 and the absolute value of each relay coefficient can also be bounded between some minimum and maximum values , the noise term in ( [ eq : input_output ] )
does not affect the dof . for detailed proof
, we refer readers to the full version of this paper .
for @xmath110 we set @xmath111 , @xmath112 , and @xmath113 .
then , from the result of lemma [ lemma : n_cancel ] , @xmath114 symbol can be cancelled at each unintended destination .
thus @xmath0 dof is achievable without symbol extension . for @xmath115 , we set @xmath116 , @xmath117 , and @xmath118 , where @xmath119 is an arbitrary integer
. then @xmath120 symbols can be cancelled .
thus the achievable dof is given by @xmath121 for @xmath122 , we set @xmath123 , @xmath124 , and @xmath125 , and @xmath126 , with @xmath127 , where @xmath119 is an arbitrary integer .
then @xmath97 symbols can be cancelled at each unintended destination because @xmath128 where we use the fact that @xmath129 .
notice that since the relays transmit linear combinations of their received signals , the @xmath39-symbol - extended channel matrix from the @xmath31-th source to the @xmath81-th destination is given by @xmath130 which means one can regard the resulting network as the @xmath0-user interference channel having the @xmath39-symbol - extended channel matrix @xmath131 . hence the remaining interference from @xmath96 unintended sources can be aligned at each destination .
we apply the interference alignment technique in @xcite to align the remaining interference and refer appendix iii in @xcite for the detailed proof .
let @xmath132 be the number of transmit symbols of the @xmath31-th source divided by @xmath39 .
then , @xmath133 similarly , we obtain @xmath134 for @xmath135 .
thus , from ( [ eq : d_1 ] ) and ( [ eq : d_i ] ) , the achievable dof is given by @xmath136 by combining the above three cases , we show that the achievable dof is @xmath137 , which completes the proof .
[ remark : converse ] if @xmath20 , then @xmath138 .
the achievability is given by theorem [ th : k_l_k ] .
the converse can be shown straightforwardly from the cut - set bound .
let us compare the resulting dof with that of the decode - and - forward ( df ) based relaying of @xcite , which is given by @xmath15 .
notice that if @xmath139 than the af based relaying provides better dof but if @xmath140 there exist values of @xmath2 for which the df based scheme is better . specifically ,
if the number of relays is less than @xmath141 or greater than @xmath142 , the af based relaying is better and the df is better otherwise .
in this section , we consider a @xmath0-user @xmath0-hop network having @xmath0 nodes in each layer and assume @xmath0 is even . in this scheme ,
relays in each layer transmit their received symbols without any modification . however , they transmit them when the channel matrix of the next hop satisfies a certain condition . since the relays may have to wait for such a channel instance for a long time , they need to store some of their past received symbols . to achieve this ,
we assume block markov coding is used at relays , i.e. , each relay receives a block of @xmath39 symbols and transmits them ( in a permuted order ) in the next @xmath39 time slots .
therefore , there will be one block delay at each hop . to simplify notations , we omit the block indices as before . in the following
we only describe how a set of transmitted symbols at any given time at the sources flows through the network .
let @xmath143 denote the vector of transmit symbols at any given time at the source nodes . furthermore , assume it is transmitted through channel matrices @xmath144 ( first hop ) through @xmath145 ( last hop ) .
we define the channel pairing rule from @xmath144 to @xmath145 so that the resulting @xmath146 becomes a scaled identity matrix .
[ fig : opportunistic_int_mit ] illustrates the basic pairing rule . applying the singular value decomposition ( svd ) ,
@xmath144 can be represented as @xmath147 , where @xmath148 consists of left singular vectors , @xmath149 is the diagonal matrix with ordered singular values , and @xmath150 consists of right singular vectors .
then we choose the next hop channel matrix to be @xmath151 , where @xmath152 is given as the singular value matrix of @xmath144 by cyclic shifting the singular values , i.e. , @xmath153 . in the same manner ,
@xmath154 can be determined from @xmath155 , and so on .
notice that these paired channels provide @xmath156 meaning @xmath0 s
d pairs can communicate concurrently without interference .
similar concepts of opportunistic channel pairing can be found in @xcite for single - hop networks and in @xcite for finite - field multi - hop networks .
let @xmath158 be a channel instance whose svd is given by @xmath159 . for @xmath160 , we define @xmath161 where @xmath162\ ] ] is the permutation matrix such that the diagonal elements of @xmath163 are equal to the cyclic shift of the diagonal elements of @xmath164 . from the definition ,
@xmath165 is given by @xmath166 , where we assume @xmath167 .
note that @xmath168 which is a scaled identity matrix denotes @xmath169 . ] . for notational simplicity , we will use the notation @xmath170 to denote @xmath171 .
for @xmath172 , we will also use @xmath173 to denote @xmath174 because @xmath175 .
[ lemma : pr_f ] suppose that the channel coefficients are i.i.d .
drawn from a continuous distribution and @xmath176 is a function of @xmath177 only .
then @xmath157 is uniquely determined by @xmath166 and @xmath178 for all @xmath179 .
let @xmath181 denote the quantized channel matrix in @xmath182 and @xmath183 denote the set of all @xmath144 whose closest point in @xmath182 is equal to @xmath181 , respectively .
we further define @xmath184 for @xmath185 . from lemma [ lemma :
pr_f ] , one can easily derive @xmath186 for all @xmath187 . for transmission
, we use the channel @xmath188 only when @xmath189 .
since @xmath190 , the probability of channel utilization can be arbitrarily close to one by setting @xmath92 and @xmath93 arbitrarily small and large , respectively , which does not affect the dof . for all @xmath181 satisfying @xmath189 , the sources transmit their messages to the nodes in the next layer through @xmath191 and the relays in the @xmath46-th layer amplify and forward their received signals to the nodes in the next layer through @xmath192 , where @xmath185 .
that is , we set @xmath193 suppose that messages are transmitted through a series of particular channel matrices @xmath144 to @xmath145 such that @xmath194 .
then from ( [ eq : input_output_vec ] ) and ( [ eq : x_m ] ) , we obtain @xmath195 where @xmath196 denotes the accumulated noise due to af relaying .
let @xmath197 and @xmath198 for @xmath160 , where @xmath199 is the quantization error matrix of @xmath200 with respect to @xmath157 . from the definition of @xmath157 , we know that @xmath201 .
then we obtain @xmath202 where @xmath203 which is the total quantization error matrix .
then the signal to interference and noise ratio ( sinr ) of the @xmath81-th destination is lower bounded by @xmath204 because we choose channel matrices satisfying @xmath212 for transmission , the relay coefficients satisfying power constraint @xmath7 can be bounded between strictly positive finite minimum and maximum values .
hence the terms @xmath213 and @xmath214 in ( [ eq : closed_form ] ) does not affect the dof .
let us now consider @xmath215 in ( [ eq : closed_form ] ) that can be represented as @xmath216 where @xmath217 denotes @xmath218-th element of @xmath166 and @xmath219 is the cofactor , which is a function of @xmath220 .
hence , for given @xmath221 to @xmath222 and @xmath223 to @xmath224 , @xmath215 becomes zero if and only if @xmath225 however this event occurs with probability @xmath226 .
therefore , this does not affect the dof and we can show dof of @xmath0 is achievable . for detailed proof , we refer readers to the full version of this paper .
the converse can be shown similarly as in remark [ remark : converse ] , which completes the proof .
g. bresler , a. parekh , and d. tse , `` the approximate capacity of the many - to - one and one - to - many gaussian interference channels , '' in _ proc .
45th annu .
allerton conf .
communication , control , and computing _
, monticello , il , sep .
2007 .
v. r. cadambe , s. a. jafar , and s. shamai ( shitz ) , `` interference alignment on the deterministic channel and application to fully connected gaussian interference networks , '' _ ieee trans .
inf . theory _ , vol .
55 , pp . 269274 , jan .
2009 .
s. mohajer , s. n. diggavi , c. fragouli , and d. tse , `` transmission techniques for relay - interference networks , '' in _ proc .
46th annu .
allerton conf .
communication , control , and computing _ , monticello , il , sep . | we study a multi - source gaussian relay network consisting of @xmath0 source destination pairs having @xmath0 unicast sessions .
we assume @xmath1 layers of relays between the sources and the destinations .
we find achievable degrees of freedom of the network .
our schemes are based on interference alignment at the transmitters and symbol extension and opportunistic interference cancellation at the relays . for @xmath0-@xmath2-@xmath0 networks ,
i.e. , @xmath3-hop network with @xmath2 relays , we show @xmath4 degrees of freedom are achievable . for @xmath0-hop networks with
@xmath0 relays in each layer , we show the full @xmath0 degrees of freedom are achievable provided that @xmath0 is even and the channel distribution satisfies a certain symmetry . |
galactic winds , driven by supernova - heated gas @xcite or agns @xcite , have been frequently observed in starburst galaxies ( e.g. , * ? ? ? * ) .
they play a significant role in enriching the intergalactic medium with metals and suppressing the star formation and growth of super massive black holes ( * ? ? ?
* and references therein ) .
they are an indispensable element in theoretical models of galaxy evolution , in which they provide the feedback mechanism that enables the models to reproduce galaxy scaling relations consistent with those observed ( e.g. , * ? ? ?
winds have also been invoked to explain a number of issues related to contemporary cosmology ( e.g. , * ? ? ?
* ) , such as the discrepancies between the observed and predicted galaxy luminosity functions ( e.g. , * ? ? ?
ultraluminous infrared galaxies ( ulirgs , defined to have @xmath2 ) are the late stages of a merger of two gas - rich galaxies , which may then form a qso ( sanders et al .
. though rare at low redshift , their co - moving density increases rapidly at high redshift @xcite .
they are hosts to intense starbursts and may dominate the star formation in the universe at @xmath3 @xcite .
detailed absorption - line studies of a large number of ulirgs have shown that winds are ubiquitous in these systems @xcite and may play a crucial role in their evolution .
our current knowledge of ulirg winds , especially their geometry , is very limited .
though previous studies of ulirgs have yielded valuable insights ( including velocity distributions , detection rates , and rough mass estimates ) , they give little or no information on the spatial variation of wind properties ( though see martin 2006 for a 1-dimensional study ) .
previous work on estimating properties of ulirg winds assumed spherical symmetry of the wind with uniform velocity and density @xcite .
since ulirg winds occur in merging systems that have more complicated dynamics than isolated disk galaxies , interaction between two galaxies can possibly affect the outflow .
for example , @xcite found that winds are detected twice as often in ulirgs as in lirgs , which suggests that the ulirg winds may have a wider opening angle compared to winds from less luminous galaxies .
if this is true , then a spherically symmetric wind model may be too simplistic to represent ulirg winds . to gain a better understanding of ulirg winds , we obtained 2-d spectra of f10565 + 2448 with the integral field unit ( ifu ) on gemini north .
outflows in our target have been studied with 1-d spectra . @xcite
have studied the velocity variations in this galaxy via long - slit spectra , and @xcite estimated the mass , energy and momentum in the wind . using the ifu data we can : 1 )
measure how extended the outflow is , 2 ) determine whether the basic model for a disk wind ( * ? ? ?
* and references therein ) applies to the late stage of galactic merger , 3 ) map the velocity structures of different phases of the outflow , and 4 ) refine estimates of the mass and energetics of the neutral phase , in which most of the mass likely resides .
f10565 + 2448 is one of the nearest and brightest starburst ulirgs selected from the survey of @xcite , with z = 0.0431 @xcite and @xmath4 @xcite .
the co redshift measured by @xcite is consistent with the optical redshift measured by @xcite .
it is classified as an galaxy by its spectral type @xcite .
however , in a more recent study , @xcite classified it as a starburst - agn composite galaxy , and the agn contribution to infrared luminosity is estimated to be 10 - 20% @xcite .
the outflow s absorption - lines structure in this galaxy shows spatial variation in long - slit spectra @xcite .
an hst image shows that our target has three nuclei , and the one we observed is by far the most luminous .
figure [ cont ] shows the hst image and a green box indicating the nucleus we observed .
we observed this object with the gemini multi - object spectrograph ( gmos ) integral field unit ( ifu ) in two - slit mode using the b600 disperser centered on 6300 , which give us a wavelength range of @xmath5 with spectral resolution of 4.26 unbinned pixels ( roughly @xmath6 ) .
each image covers a science field of @xmath7 , which at this redshift is about @xmath8 kpc , with 1000 lenslets .
there were two dithering positions with offsets of 15 from the center in the east - west direction . with eight exposures of 1800 seconds each , a total of four hours worth of data was collected .
seven frames were taken on the night of 2007 feb 12 ( ut ) and one was taken on 2007 jul 4 ( ut ) .
the seeing was variable within the range of 10 - 13 .
the data were reduced using the gemini iraf package ( version 1.8 ) following the standard steps : overscan subtraction , bias subtraction , flat - fielding , wavelength calibration , sky subtraction , flux calibration , and spectra extraction .
the reduced 2-d images were assembled into a 3-d data cube with coordinates ( x , y , @xmath9 ) using the gfcube routine . for each data cube ,
differential atmospheric refraction was corrected by shifting the image slices at each wavelength to keep the centroid of the nucleus constant . after centering and combining all the cubes , and then trimming off the uneven edges , the total science field that was used for analysis is @xmath10 or @xmath11 kpc .
the original resolution of the images was 02 per pixel . to get better signal to noise
we binned the data cube to 06 per pixel , and since this is smaller than the seeing , we did not lose significant spatial information .
the final combined data cube was sliced into 143 1-d spectra that can be used by our absorption - line fitting routine . in our final data cube , there appears to be a resolution offset between the two sides of the field of view . after working with the gemini helpdesk
, we learned that there is a known resolution difference between the two pseudo - slits that is probably flexure dependent .
since the reason for the offset is still unknown there is no standard procedure to fix the problem . after inspecting the reduced data
, we determined that we can make a meaningful analysis without any correction .
this resolution difference in the wavelength direction can be up to 10 km / s which could possibly affect our spectral fitting , however , we have checked the fitted parameters and found no clear systematic effects between the two pseudo slits .
methods for analyzing superwinds using spectroscopy utilize both emission and absorption lines . however emission lines can be contaminated by light from star forming regions , and they have an observational degeneracy between infall and outflows .
absorption lines have two advantages in the case of ulirgs .
first , emission lines in ulirgs are dominated by rest - frame starburst component ( e.g. * ? ? ?
* ; * ? ? ?
* ) , while the absorption - line rest - frame component is typically swamped by the outflow component @xcite .
this is at least partly due to the different density dependence of emission and absorption - line strengths .
infalls and outflows sources can also be easily distinguished for absorption lines , since an absorbing cloud has to be between the observer and the light source . in emission lines , on the other hand
, blueshifted gas could be inflowing or outflowing , depending on the ( unknown ) geometry .
@xcite developed software for detailed analysis of the outflows via fitting the na i d @xmath12 absorption lines . because of its low ionization energy , na i
d probes the neutral gas , which carries significant amounts of mass and energy and is easily detectable in these galaxies .
we used this software to analyze the properties of the wind in f10565 + 2448 .
we use the fitting routine , nafit , developed by @xcite to analyze the na i d feature .
this routine fits multiple velocity components to the na i d absorption lines by varying the covering factor , optical depth , line width ( or more specifically the doppler parameter b = @xmath13 = fwhm/ @xmath14 $ ] ) and central wavelength of each component , and minimizing @xmath15 using a levenberg - marquardt optimizer . for each line , assuming a continuum level of unity , the intensity of the absorption line is given by @xmath16 where @xmath17 is the covering factor , and @xmath18 is the optical depth . assuming a maxwellian velocity distribution , @xmath18 can be expressed as a gaussian : @xmath19 where and @xmath20 and @xmath21 are the central optical depth and central wavelength in the line and b is the width of the line . as the na i d
feature is a doublet , and the ratio of the optical depth of the two lines is determined by atomic physics , the central optical depth of the stronger line is always twice that of the weaker line i.e. , @xmath22 . at any particular wavelength ,
@xmath23 where @xmath24 is given by equation [ tau ] for each line .
we also assume an identical covering factor for both lines in a doublet . with this assumption , both optical depth and covering factor can be constrained simultaneously . for a detailed discussion on accounting for blended lines and overlapping absorbers ,
see @xcite .
we wrote a wrapper to this fitting software to fit all of our spectra automatically .
starting at the center of the galaxy , where signal - to - noise is the highest , we step outward in a spiral pattern , using the output parameters of the previous fit as the initial guess for the next spectrum .
fig [ snimg ] is a signal - to - noise map of the continuum near the na i d line , in units of s / n per unit spectral resolution element .
the s / n is as high as 50 at the center of the nucleus and declines to as low as 2 or 3 at the edges of the fov .
as shown in fig [ 12comp ] , at the continuum peak , the absorption line is best fitted by two velocity components .
( note : each velocity component contains a pair of doublet lines ) .
the one component fit can reproduce the basic shape of the line but leaves significant residuals .
a two component fit lowered the residuals significantly , and three or more components over - constrains the fit .
as we move outward to regions with lower signal - to - noise , we find spectra that can be fit by one component equally well as by two .
however , switching between one and two component fits within the same data cube introduces artificial patterns to our velocity maps .
therefore , we first perform two - component fits for all spectra ; we then visually inspect the fits and remove any components that have negligible contribution to the equivalent width of the line .
examples of significant vs. insignificant components are shown in fig [ sigfit ] .
some spectra show the he i @xmath25 emission line which appears immediately blueward of the na i absorption lines .
we include this emission line in our fit , where necessary , to remove its effect on our absorption - line fitting . for most of the spectra ,
however , the he i emission is weak enough that it blends in with the noise and can be safely ignored .
figure [ hefit ] shows examples of he i line fitting .
all parameters are allowed to vary freely during the fit .
upper and lower limits of all parameters are the same for all spectra except for the central wavelength which is @xmath26 of the previous fit .
the upper limit of the optical depth @xmath18 is set to 5 since our fits are insensitive to optical depths beyond that .
the covering factor is constrained to be 0 to 1 . to ensure that we are not extracting velocity dispersion widths smaller than the spectral resolution , the lower limit of the doppler parameter is set to the spectroscopic resolution of @xmath27 km / s determined from the line width of the calibration lamp lines .
the fitted velocity widths are typically well above this lower limit .
it should also be noted that @xcite have shown that not de - convolving the instrumental resolution before the fitting does not change the result significantly . the upper limit of the velocity width is set to @xmath28 km / s and no spectra came close to exceeding this limit . we have checked our monte carlo simulation outputs for possibilities of fitting degeneracies . since this step essentially fits slightly modified versions of the same spectrum 1000 times , degeneracies for any parameters should show up as a bimodal or more complicated distribution in the fitting results .
we find that almost all of the fitted parameters show a roughly gaussian distribution .
exceptions occur mostly for the lower s / n spectra around the edges that are later discarded anyway due to their low quality .
we fit 5 emission lines ( [ oi ] 6300 and 6364 , h@xmath1 , and [ ] 6548 and 6583 ) using a suite of routines that simultaneously estimates the stellar continuum and fits all emission lines at once , with the widths and redshifts constrained to be the same for each emission line .
the stellar continuum , minus emission lines , was estimated using the stellar synthesis models of @xcite .
the emission lines were fit with two gaussians , one narrow and one broad ; the broad component was discarded near the edges of the data , where the signal - to - noise was too low .
the emission - line fitting routines are described further in rupke et al .
( 2010 , in prep . )
figure [ cont ] is the hst / acs two color image of our target , with f814w shown in red and f435w shown in green .
the green box on the lower right nucleus indicates the gmos field of view , and the continuum image of our data is shown in the lower left .
the continuum flux map of our target is integrated over 6400 - 6500 . as mentioned in section 2 ,
the pixels are binned @xmath29 to get better signal to noise .
the h@xmath1 flux map is shown in the top two panels of figure [ em_lines ] .
we ve applied a global balmer - decrement - determined extinction correction to the fluxes , e(@xmath30 ) = 1.75 @xcite , which corresponds to @xmath31 .
for both components , the h@xmath1 flux peaks at the same pixel as the continuum flux map .
the bottom two panels show the [ ] /h@xmath1 flux ratio map for both components . for the narrow component , the line ratio is fairly uniform and symmetric around the center and has an average value of 0.5 , consistent with star formation @xcite .
away from the nucleus , however , the ratio increases .
the broad component has a much wider range of line ratios and a peak of 1.8 appears in the south and northwest .
using the h@xmath1 flux from the emission line fitting we can estimate a star formation rate ( sfr ) . integrating over the h@xmath1 flux map we obtained a total flux of f(h@xmath1 ) = 4.08 @xmath32 erg s@xmath33@xmath34 . using ned wright s javascript cosmology calculator @xcite , assuming @xmath35 km s@xmath33mpc@xmath33 , @xmath36 and @xmath37
, we get a luminosity distance of @xmath38 mpc , and therefore l(h@xmath1 ) = 1.7 @xmath39 erg / s .
sfr can be estimated using the following equations @xcite : @xmath40 @xmath41 the sfr estimated from l(h@xmath1 ) is about 136 @xmath42 yr@xmath33 , and 190 @xmath42 yr@xmath33 from @xmath43 .
the sfr estimated from h@xmath1 flux is slightly smaller as expected since the extinction from @xcite is a lower limit .
figure [ velmaps ] shows the velocity map of the na absorption lines and emission lines respectively .
( note : since the signal to noise is very low at the edges of the fov , we did an additional @xmath44 binning at the four corners so we can constrain the parameters better .
this only applies to the absorption - line maps and not the emission - line maps .
this additional binning shows up as four large pixels , which are 9 times larger than the rest of the pixels , at the four corners of all the absorption - line velocity and velocity dispersion maps . )
velocities are calculated with respect to a fixed redshfit of @xmath45 . the narrow emission line map , which we interpret as systemic , is blueshifted in the east and redshifted in the west and clearly shows rotation ( see section 5 for more discussion ) .
the pa of the major axis according to @xcite , 100@xmath46 east of north , is plotted as red lines on the absorption - lines and narrow emission - line velocity maps as a reference .
most of the northern half of the broad line emission map has values very close to systemic , but the southern half shows significant blue shifts .
the absorption lines are separated into higher - outflow - velocity ( blue ) and lower - velocity ( red ) components .
both components are generally blue shifted with respect to systemic and their velocity gradients are east - west oriented , roughly aligned with the major axis . figure [ vdisp ] shows the velocity dispersion , or full width at half maximum ( fwhm ) , maps of the emission and absorption lines respectively .
both narrow and broad emission lines have a high velocity dispersion region in the southeast , roughly in the same region as the peak of [ ] /h@xmath1 .
the red absorption component has a typical fwhm of roughly 200 km s@xmath33 , while the bluer component has a higher typical fwhm at about 300 km s@xmath33 .
the systemic emission - line fwhm increases away from the nucleus from about 150 to 350 km s@xmath33 .
nafit calculates the error in each fitted parameter using monte carlo simulations described in detail in @xcite .
errors in the velocities are 5 - 10 km s@xmath33 close to the center and remain @xmath47 30 km s@xmath33 for most of the pixels , although for some pixels close to the edge the error can be @xmath48 50 km s@xmath33 .
fwhm errors start from @xmath47 10 km s@xmath33 in the center , remain below 50 km s@xmath33 for most pixels , and increase up to @xmath48 80 km s@xmath33 around the edges .
other than the random errors due to noise , the systematic effect between the two pseudo slits mentioned in 2 may also cause errors .
the two pseudo slits are oriented north - south with respect to each other .
judging from the images of the fitted parameters , we can see that there is no clear systematic offset in the north - south direction ( the direction of the systematic offset ) .
the velocity gradient is oriented in the east - west direction , and from the arc lamp measurements we find that the resolution offset is small ( about 10 km s@xmath33 ) compared to the velocity gradient observed .
therefore we can safely conclude that any errors caused by the systematic offset have a negligible effect on our analysis .
our target is known to be dusty and we can test whether the neutral gas absorption corresponds to the dust features .
a map of the na i d line equivalent width is shown in the right panel of fig [ color_dust ] .
the equivalent width shown here is the total equivalent width of both velocity components .
a map of the dust features is shown in the sdss @xmath49 ratio map in left panel of fig [ color_dust ] .
we are using @xmath49 ratio as a proxy for e(@xmath30 ) and assuming that the stellar population does not vary too much across the field of view .
a lower @xmath49 ratio indicates a redder color and corresponds to the easily visible dust filaments in the high resolution hst image ( fig [ cont ] ) .
the sdss @xmath50 and @xmath51 images are rotated and resampled to match the pa and resolution of our data .
a pixel to pixel comparison plot of @xmath49 ratio and equivalent width images is shown in fig [ gi_eqw ] .
the na i d line strength correlates well with the dust reddening .
we compared our ifu data with the freshly analyzed long - slit observations of our target along a position angle of 53 degrees @xcite .
the long slit data probe further from the nucleus , and show the presence of a neutral ( ionized ) wind up to @xmath52 ( 10 ) kpc from the nucleus , as well as regions of large line broadening and shock excitation that extend 10 kpc from the nucleus .
this is consistent with similar results from @xcite along a pa of 109 degrees .
quantities calculated from @xcite long - slit data , including outflow velocities , velocity dispersions , h@xmath1 fluxes , and [ ] /h@xmath1 ratios , agree with our measurements from the ifu data .
the narrow line emission shows a clear rotation signature with blue shifted lines in the east and redshifted lines in the west , which is consistent with long - slit h@xmath1 observations by @xcite and @xcite . within errors ,
this is also consistent with the co presented in @xcite .
our velocities ranged from -122 to 35 km s@xmath33 , while @xcite measured -70 to 70 km s@xmath33 , relative to central velocity , in the molecular disk .
in general , the narrow emission - line velocities follow the rotation of the molecular disk closely .
the small velocity shift across the nucleus is consistent with the galaxy being almost face on in our line of sight ( downes & solomon 1998 estimated @xmath51 = 20 degrees ) .
based on the velocity map , the outflow extends at least 6 kpc in diameter across the center of the galaxy , confirming previous results @xcite .
according to the h@xmath1 and co flux distributions , the launching region of the wind is roughly the inner 1 kpc of the galaxy where the starburst and the molecular gas are concentrated . since the h@xmath1 flux peak is the same as the continuum peak , we define the launching region as the area within 1 kpc of the continuum peak .
thus , the wind extends at least 2 kpc away from the source region in all directions ; this distance is a lower limit if we assume extensions along the line - of - sight .
we see evidence of asymmetric outflow in the broad emission - line and absorption - line velocity maps as mentioned in the previous section . in principle , red shifted emission lines can be observed from wind receding on the opposite side , but they are not observed most likely due to dust obscuration , which is consistent with other ulirgs observations @xcite .
the general pattern of the broad emission - line velocity clearly differs from that of the absorption lines .
the broad emission lines only appear to be blue shifted in the southern region while the absorption lines are blue shifted all across the map .
the warm ionized gas that produces the blue shifted emission lines we observed arises at boundaries between the wind fluid and ambient gas .
though the relationship between velocities of absorbing - line clouds and line - emitting gas is not well known , previous studies ( e.g. , * ? ? ?
* ; * ? ? ?
* ) have suggested that they may have different velocity structures , which is consistent with our observations .
the surprising result from the absorption - line velocity maps is that the velocity gradient is opposite to that of the systemic rotation and aligned with the major axis .
in general , the narrow emission - line velocity map is blue shifted in the east and red shifted in the west .
the outflow velocity maps are less blue shifted in the east and more blue shifted in the west .
@xcite observed our target with a long slit going through the continuum peak of the nucleus we observed , and a pa of @xmath53 degrees .
their data suggested that the velocity of the neutral outflow is roughly constant across the entire galaxy , which implies an outflow diverging from a central starburst and slowing down with increasing height to conserve angular momentum .
however , in our observation we do see a significant velocity jump in the absorption lines .
outflows with significant velocity shear along the major axis may be at a low scale height and therefore retain the rotation velocity of the host galaxy , but in this scenario the velocity gradient should be in the same direction as the systemic gradient while the velocity gradient we observed is in the opposite direction .
one can compare the data to a simple toy model .
consider a galactic wind which consists of a gas disk and central starburst from which the wind is launched , and an outflow that extends perpendicular to the disk and away from the center in a conical shape ( see , e.g. , martin 2006 for graphic illustrations ) .
according to this model , an outflow that conserves angular momentum shows very little velocity gradient along the major axis , but may show a velocity jump along the minor axis if the galaxy is tilted with respect to the observer .
we do not see a significant velocity gradient in the minor axis direction . however ,
since the central disk is almost face on , a velocity gradient along the minor axis due to projection effects may not be detectable in our case . in the context of this simple model
, one could observe a major axis gradient if the entrained outflowing material maintained the rotation it had prior to entrainment .
however , this is inconsistent with the fact that the observed gradient is opposite to disk rotation .
we thus conclude that the simple model does not adequately explain our observations .
a wind perpendicular to the central disk is still a possibility , but another ingredient is necessary to explain the major axis velocity gradient .
another possible explanation of the observed outflow velocity is an uneven power source .
if there is a stronger starburst in the west , and this is accelerating the wind at a faster rate , the wind may appear to be counter rotating with respect to the galaxy .
but this explanation is inconsistent with the distribution of star formation ( h@xmath1 flux ) and the gas reservoir ( co flux ) , which are both centrally concentrated .
our best explanation is that the velocity difference is caused by interaction between the wind and ambient gas .
the broad emission component , with high velocity dispersion and [ ] /h@xmath1 ratio , is detected up to 2.5 kpc from the nucleus , indicating a widespread shocking of interstellar gas .
@xcite found a broad emission - line component in m82 outflow with increasing line width away from the nucleus ; they attribute this to turbulence as the wind shocks and mixes with the surrounding gas .
if our broad emission - line component is of the same nature , varying densities in the ambient gas may be the cause of velocity asymmetry .
more evidence that indicates shocks caused by winds can be seen in the broad emission - line [ ] /h@xmath1 ratios which ranges from @xmath540.8 to @xmath542.0 .
such high ratios can only be explained by high velocity shocks @xcite or an agn . since starburst dominates the luminosity in f10565 + 2448 @xcite , high velocity shocks are the most plausible explanation .
previous work has used the correlation between ionization and velocity dispersion as evidence of shock induced ionization ( e.g. * ? ? ?
* ; * ? ? ?
our [ ] /h@xmath1 and velocity dispersion maps show that higher ionization corresponds to higher velocity dispersion ( bottom left panels of fig [ em_lines ] , [ vdisp ] ) , and supports a scenario in which outflowing gas interacts with and shocks the surrounding materials .
variation of density in surrounding gas can be seen in a localized high ionization region found in the southeast of the broad emission line [ ] /h@xmath1 map ( fig [ em_lines ] ) .
this region corresponds to a high velocity dispersion region for both the broad and narrow emission lines ( fig [ vdisp ] ) , which suggest that the gas density in that region is higher and shock induced ionization dominates the emission lines .
since this region coincides with lower outflow velocities in its neutral phase , the higher density interstellar medium may be decelerating the outflow gas .
@xcite showed that ulirg winds have wider opening angles ( @xmath55 steradians ) than lower luminosity galaxies based on the detection rate of winds in ulirgs , which is almost twice as high as in less luminous galaxies .
the inconsistency between our outflow velocity maps and a basic disk wind model supports the hypothesis that the morphology of ulirg winds may be different from winds of disk galaxies .
the asymmetry of the outflow velocity maps may be a result of disturbance in the ism due to merging .
thus , a merger may affect the dynamics of the wind , causing it to deviate from a simple disk wind structure .
the above discussion assumes that the outflowing material comes from the starbursting disk .
it is possible that the outflow may be dominated by gas entrained outside of the disk , which may have a different velocity pattern due to the complex merger dynamics .
for instance , the edges of our systemic velocity map show deviations from simple rotation .
however , we show below that the outflowing gas mass is high , suggesting that an origin in the massive central gas reservoir is most plausible .
furthermore , the outflow is dusty , implying that there is neutral , perhaps even molecular , gas in the wind , most likely originating in the dusty molecular disk .
an outstanding issue in galactic wind research is the relationship of different phases of the wind to each other .
ionized gas has sometimes been seen in outflow in concert with neutral gas in starburst - powered ulirg winds @xcite .
globally , ulirg winds are also thought to be dusty , due to the strong positive correlation between equivalent width of the na i d line and e(b - v ) in lirgs and ulirgs .
using our data , we can convincingly demonstrate that the ionized gas is outflowing on kiloparsec scales , with high velocities ( exceeding 600 km s@xmath33 in some locations ) , large linewidths ( up to 600 km s@xmath33 ) , and strong collisional excitation .
@xcite have detected blueshifted ionized gas in ulirgs with agns using mid - infrared emission lines .
we have shown that starburst ulirgs can also possess ionized outflows . as discussed in 5.1 , the correlation with
the neutral gas is weak , however , suggesting that the relationship between the ionized and neutral gas is complex .
observations of more resolved sources may be necessary to understand the physical link between the two .
our data also show , for the first time , a strong spatial correlation between dust reddening and nai d equivalent width , again on kiloparsec scales .
this robustly confirms the physical coincidence of the neutral gas and dust in these outflows . by inference
, it also suggests that the filamentary dust features seen in hst images of f10565 + 2448 and other ulirgs are in many cases instances of outflowing dust and gas .
this result raises again the tantalizing possibility that starburst- and agn - driven winds are responsible for removing the cocoon of dust around obscured agn in mergers .
mid - infrared spectra of ulirgs reveal deeply - absorbed continua @xcite ; these large optical depths have been hypothesized to hide nascent quasars @xcite .
outflows like the one observed in this source may be the mechanism by which this dust is removed and a quasar is born
. we can refine the estimation of mass and kinetic energy in the neutral wind using a wind model that depends on the physical parameter outputs of our fitting .
we choose a spherical thin shell instantaneous outflow similar to the one used by @xcite .
the difference is that @xcite assumed spherically symmetric velocity and density , but we have information on the spatial variations of these quantities .
an alternative to the thin shell model is an outflow with finite thickness and velocity that depends on radius . in @xcite and @xcite ,
the thick shell outflow has @xmath56 , where @xmath57 is the ( spherical ) radial distance from the nucleus and @xmath58 is the outflow velocity . because we observe the galaxy in a face - on orientation , we can not easily probe variations in velocity with @xmath57 , so we assume a single , characteristic value for @xmath57 but let @xmath58 vary across the outflowing shell according to the velocity output of nafit .
we have de - projected the velocities from observed to purely radial velocities .
the instantaneous mass outflow rate , average outflow rate and total mass are : @xmath59 @xmath60 @xmath61 where @xmath57 is the radius of the thin shell , theta and phi are the usual spherical angular coordinates , @xmath62 is the thickness of the shell ( @xmath63 ) , @xmath64 is the mean molecular weight , @xmath65 is the mass of a proton , and @xmath66 is the column density of hydrogen .
for all our calculations we choose @xmath67 kpc to be consistent with @xcite , and @xmath68 . using the mass outflow rate
we can calculate the energy and momenta outflow rates ( see rupke et al .
2005b ) , the values are summarized in table [ outflowrates ] .
the column density of hydrogen , n(h ) , can be estimated from n(nai ) given a metallicity and ionization fraction : @xmath69 where @xmath70 , @xmath71 = -5.69 $ ] which we set to solar value , and @xmath72 - log[n(na)/n(h)_{gas } ] = -0.95 $ ] is the assumed depletion .
the values that we calculated for mass , energy , momentum and their outflow rates are within a factor of @xmath73 of the values in @xcite .
the new mass , energy and momentum are smaller than what is predicted from long slit data , while the outflow rates are larger . with the new spatial information on velocity and column densities
, the new values should be more accurate .
the thin shell geometry that we assumed is , however , uncertain .
theory and simulation predicts a thin shell of warm gas behind the forward shock of the wind ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , but it is unclear whether it is accompanied by neutral gas .
since we have limited information on the radial depth of the wind , we can not rule out a thick outflowing shell in which neutral clouds are continuously entrained in the wind . switching from a thin shell to thick shell model
will decrease our measured mass .
furthermore we are only considering the part of the wind where we have reliable spectral fitting ; the actual wind probably extends further than what we observed and carries more mass . in some optically thick cases ,
our fitting routine finds only a lower limit for @xmath18 for one of the velocity components , which is set to 5 in our fits .
this is applicable to @xmath74% of the spectra used for this section s calculations , and may cause us to underestimate the mass .
we check the extend of this effect by comparing our values to those calculated by using only one velocity component fit throughout .
when only one component is used for fitting , only @xmath75% of the fitted @xmath18 used for calculation are at the lower limit .
the estimated mass and energy are slightly smaller than , but within a factor of 2 of the values presented in table [ outflowrates ] . from the present data alone , we can not rule out the possibility that optically thick lines cause a significant underestimate of the mass in the wind . however , we believe it is more likely that our mass estimates are not biased in this way because : ( 1 ) the saturated components in the current data arise primarily at low signal - to - noise ; ( 2 ) the one - component fits , which have more sensitivity to the overall shape of the lines in the low s / n region did not give a drastically different mass estimate .
higher s / n data will likely resolve this issue and better constrain the optical depths of each component .
since our observation only trace the nai d clouds , it is possible that we underestimated the total mass of the wind by not considering the other phases .
@xcite compared ratio of cold , neutral gas outflow rates to the expected mass injection rated from the hot gas created by supernovae .
the ratios are found to be close to unity on average but it can vary from 0.001 to 10 for individual galaxies . on the other hand , theory suggests that the entrained clouds dominate the mass budget of the wind ( e.g. * ? ? ?
* ) , and we suspect that the nai d clouds are the entrained gas @xcite .
the conversion factor used to convert the neutral na column density to that of the total na has also been shown to be uncertain @xcite . however
, the total mass in the cold gas can not be more than an order of magnitude larger than our calculated value , otherwise the mass ratio between cold and hot gas will be too large @xcite . within the above constraints , the present data put on firmer footing the coarse mass outflow rates measured by @xcite .
it is clear that these outflows are massive and energetic .
the neutral outflow rate in this particular system is comparable to the star formation rate .
however , this particular system has a higher outflow rate than the average value in ulirgs @xcite . further spatially resolved and multiphase observations are necessary to understand the exact relationship between mass outflow rate and star formation rate in galactic superwinds .
table [ outflowrates ] + outflow characteristics .mass , energy , momentum and the outflow rates in logarithmic values .
column ( 1 ) the log of total quantity of mass , energy and momentum .
column ( 2 ) the log of instantaneous outflow rate . column
( 3 ) the log of average outflow rate .
the units are the following : mass - @xmath76 , mass outflow rates - @xmath76 yr@xmath33 , energy - erg , energy outflow rates - erg s@xmath33 , momentum - dyne s , momentum outflow rates - dyne . [ cols="^,^,^,^",options="header " , ]
we have presented gemini integral - field data of a galactic wind in the starburst ulirg f10565 + 2448 to study the structure of the wind in detail.we fitted the na i d absorption lines and the h@xmath1 and [ ] emission lines across the @xmath77 kpc field of view using multiple velocity components .
the systemic rotation velocity determined from the narrow emission lines is consistent with the molecular gas disk rotation .
the absorption - line gas velocities shows asymmetric outflows with velocity gradients along the major axis opposite to that of the systemic rotation .
we see widespread evidence of shock ionization through collisional excitation of the gas and large velocity dispersions , with particularly strong shocks observed in the southeast corresponding to the region with lower absorption - line outflow velocities .
this supports the notion of gas deceleration in the east , which should be accompanied by shocking .
our results strengthen the hypothesis that ulirg winds may differ in structure from those in the more quiescent galaxies .
we observed both neutral and ionized outflowing gas .
the correlation between the outflow velocities of these two phases is weak , suggesting a complex relationship between the two .
we found a spatial correlation between dusty region and high na i d equivalent width region , confirming spatial coincidence of dust and neutral entrained outflow material .
the presence of dust in the outflow raises the possibility that winds may be responsible for removing dust around obscured agn and revealing a bright quasar .
we calculated the mass , energy and momentum outflow rate assuming a spherical thin - shell outflow .
the numbers that we obtained are in general agreement with those in @xcite . however , by studying the structure of winds in other starburst mergers , we should be able to improve on the current results .
first , by studying edge - on disks we will be able to better constrain the opening angles of these winds .
second , by studying a sample of winds , we will place better constraints on the relationship between the ionized and neutral phases of the wind , and whether asymmetric velocity structures are common .
finally , and perhaps most importantly , we must compare observations to detailed simulations of mergers that include realistic wind prescriptions such simulations do not yet exist , but are necessary for understanding detailed data sets like the present one .
the authors thank jabran zahid for sharing routines that were the framework for our stellar and emission - line fitting software .
this paper used data from the nasa / esa hubble space telescope , and obtained this data from the nasa / esa hubble space telescope , and obtained fro m the hubble legacy archive , which is a collaboration between the space telescope science institute ( stsci / nasa ) , the space telescope european coordinating facility ( st - ecf / esa ) and the canadian astronomy data centre ( cadc / nrc / csa ) .
this paper used data from the sloan digital sky survey ( sdss ) .
funding for the sloan digital sky survey ( sdss ) has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the u.s .
department of energy , the japanese monbukagakusho , and the max planck society .
dsnr is supported by a university of hawaii start - up grant to lisa kewley , and by nsf career grant ast07 - 48559 .
this work is based on observations obtained at the gemini observatory , which is operated by the association of universities for research in astronomy , inc .
, under a cooperative agreement with the nsf on behalf of the gemini partnership . | neutral outflows have been detected in many ultraluminous infrared galaxies ( ulirgs ) via the na i d @xmath0 absorption - line doublet . for the first time , we have mapped and analyzed the 2-d kinematics of a cool neutral outflow in a ulirg , f10565 + 2448 , using the integral field unit ( ifu ) on gemini north to observe the na i d feature . at the same time we have mapped the ionized outflow with the [ ] and h@xmath1 emission lines .
we find a systemic rotation curve that is consistent with the rotation of the molecular disk determined from previous co observations .
the absorption lines show evidence of a nuclear outflow with a radial extent of at least 3 kpc , consistent with previous observations .
the strength of the na i d lines have a strong , spatially resolved correlation with reddening , suggesting that dust is present in the outflow .
surprisingly , the outflow velocities of the neutral gas show a strong asymmetry in the form of a major - axis gradient that is opposite in sign to disk rotation .
this is inconsistent with entrained material rotating along with the galaxy or with a tilted minor - axis outflow .
we hypothesize that this unusual behavior is due to an asymmetry in the distribution of the ambient gas .
we also see evidence of asymmetric ionized outflow in the emission - line velocity map , which appear to be decoupled from the neutral outflow .
our results strengthen the hypothesis that ulirg outflows differ in morphology from those in more quiescent disk galaxies . _
subject keywords : _ galaxies : individual ( f10565 + 2448 ) galaxies : interactions galaxies : kinematics and dynamics |
the insulating parent compounds of cuprate superconductors are known to have an antiferromagnetic ( af ) order of cu@xmath4 spins .
because of the quasi - two - dimensionality of the cu@xmath4 spin structure , these materials can show rather strong magnon heat transport at relatively high temperatures.@xcite however , the magnon transport seems to be negligible at low temperatures , probably due to the anisotropy gap of the spin spectrum .
the external magnetic field can strongly suppress or close the gap at some critical fields , usually accompanied with the spin - flop transitions .
it was therefore expected to observe the magnon heat transport in magnetic field.@xcite in last several years , a few works on this topic have been performed on parent compounds of electron - doped cuprate superconductors , pr@xmath5la@xmath6cuo@xmath2 ( plco ) and nd@xmath1cuo@xmath2 ( nco).@xcite the low-@xmath7 heat transport of plco seemd to be quite simple and indicated that some paramagnetic moments such as free " cu@xmath4 spins can scatter phonons and lead to non - negligible magnetic - field dependence of phononic thermal conductivity.@xcite it is notable that the plco data did not indicate the possibility of magnon heat transport , although the spin - flop transition can happen for not strong in - plane fields . on the other hand ,
the nco results are more complicated because of the complexity of magnetism caused by the nd@xmath3 ions.@xcite it is known that in nco the cu@xmath4 spins order antiferromagnetically below @xmath8 k with a noncollinear magnetic structure,@xcite that is , the cu spins in the same cuo@xmath1 plane are antiferromagnetically ordered and those in the adjacent planes along the @xmath9 axis are perpendicular to each other .
the moment of nd@xmath3 ion is enhanced drastically with lowering temperature and a long - range af order develops below @xmath10 1.5 k with the same noncollinear spin structure as cu@xmath4 spins.@xcite when the magnetic field is applied in the cuo@xmath1 plane , the cu@xmath4 spins can re - orientate and enter a spin - flopped state.@xcite the transition fields are reported to be 4.5 t and 0.75 t for @xmath11 and @xmath12 $ ] , respectively.@xcite in addition , the strong coupling between nd@xmath3 spins and cu@xmath4 spins , which was found to be about 4 t,@xcite drives nd@xmath3 spins to rotate together with cu@xmath4 spins , keeping their relative orientation unchanged , and forms a three - layer unit consisting of one cu@xmath4 layer sandwiched between two nd@xmath3 layers as a whole . for this reason ,
the nd@xmath3 spins are generally considered to change together with cu@xmath4 sublattice under the influence of magnetic field . however , at low temperatures ( @xmath13 1.5 k ) when the af order of nd@xmath3 spins is formed , the magnetic structures and field - induced transitions of nd@xmath3 sublattice are still unclear since all the earlier neutron scattering in applied field were carried out at relatively high temperatures .
two earlier studies on the low-@xmath7 thermal conductivity of nco have revealed strong magnetic - field dependence of @xmath0 , which was believed to be associated with the spin flop of cu@xmath4 spins.@xcite an additional conduction channel of nd@xmath3 magnons was supposed to be the reason for the enhancement of @xmath0 in high fields , considering the reorientation of nd@xmath3 spins under nd@xmath3-cu@xmath4 interaction
. however , there are some quite unclear issues calling for more careful investigations .
first , it is apparently questionable in ref . to attribute the large increase of @xmath0 above some transition fields to the nd@xmath3 magnons at temperatures above 2 k , where the long - range ordering of nd@xmath3 spins has been lost . in particular
, the field - induced enhancement of @xmath0 is even much larger at 5 k than that at 2 k.@xcite in that work , more direct evidence for nd@xmath3 magnon transport , which should be detected below the nel transition of nd@xmath3 spins , had not been investigated .
another work performing measurements down to milli - kelvin temperatures , however , did not report the magnetic - field dependent data,@xcite which is indispensable for showing the relationship between heat transport and the transition of magnetic structure . second
, although the low - energy magnons can be significantly excited at the spin - flop transition and may contribute to carrying heat , because of the closure of the gap in magnon spectra , further increasing magnetic field will always shift the magnon dispersion upward and finally suppress the low - energy magnon excitations.@xcite so it is odd to assume that the magnon heat transport induced by the spin flop transition can be still active in the high - field limit.@xcite third , it was assumed that the nd@xmath3 spin lattice should rotate together with cu@xmath4 sublattice in the in - plane field .
however , it is notable that the nd@xmath3-nd@xmath3 interaction can be enhanced significantly at very low temperatures due to the increased magnitude of nd@xmath3 moment , which means that the nd@xmath3 sublattice could re - orientate independently . in this work ,
we study the temperature and magnetic - field dependencies of @xmath0 of high - quality nco single crystals in great detail to investigate the role of magnons in the low-@xmath7 heat transport .
our results on @xmath14 isotherms demonstrate that the phonon conductivity is strongly dependent on the magnetic field , and the nd@xmath3 magnon can play a dual role in the transport as the heat carriers or the phonon scatterers .
to be exact , the nd@xmath3 magnons mainly scatter phonons at the critical region ( 1@xmath102 k ) of nel transition in zero field , while they can significantly transport heat in the spin - flopped state induced by the in - plane field .
intriguingly , the drastic behaviors of @xmath14 under the in - plane field reveal the spin re - orientation transitions of nd@xmath3 sublattice at sub - kelvin temperatures .
high - quality nd@xmath1cuo@xmath2 single crystals are grown by using the slow cooling method with cuo@xmath1 as a self flux .
the plate - like samples for transport measurements are confirmed to grow along the @xmath15-plane by the x - ray back - reflection laue photographs .
the dimensions along the @xmath9 axis of nco crystals are so small that the measurement with heat current along the @xmath9 axis is difficult to carry out .
the thermal conductivity is therefore measured only along the @xmath16 axis by using a conventional steady - state technique and two different processes : ( i ) using a one heater , two thermometers " technique in a @xmath17he refrigerator and a 14 t magnet at temperature region of 0.3 8 k ; ( ii ) using a chromel - constantan thermocouple in a pulse - tube refrigerator for the zero - field data above 4 k.@xcite
-axis thermal conductivity of nd@xmath1cuo@xmath2 single crystal in the zero field , compared with the results reported by jin _ et al._@xcite and li _
et al._@xcite inset : the ratio of the phonon mean free path @xmath18 to the averaged sample width @xmath19 in the zero field.,width=264 ] figure 1 shows the temperature dependence of thermal conductivity of a nco single crystal in zero field , together shown are data from some other groups.@xcite the large peak at 20 k is apparently the so - called phonon peak in insulators.@xcite it is known that the magnitude of the phonon peak is dominated by the phonon scattering by the crystal defects and impurities and is therefore a good characterization for the crystal quality.@xcite it is clear that the phonon peak of our crystal is almost the same as the data obtained on a nco crystal grown using the floating - zone method,@xcite suggesting the high quality of our nco crystal . although the high - quality nco had not been measured down to very low temperatures in the earlier work,@xcite our crystal can alternatively demonstrate the intrinsic low-@xmath7 heat transport of this compound .
below 10 k , the thermal conductivity decreases quickly with lowering temperature , like usual insulators , but there is an obvious variation in slope of the @xmath20 curve around 1.5 k , which is apparently related to the af ordering of nd@xmath3 ions.@xcite however , only from the zero - field data , it is not clear whether the slope change is caused by an enhanced magnon scattering on phonons at the critical region of af phase transition or an appearance of the magnon heat transport below the transition . the magnetic - field dependence of @xmath0 is known to be useful to clarify this issue . before analyzing the @xmath14 data , we can make an estimation of the phonon mean free path ( @xmath18 ) at very low temperatures assuming that the magnons are not acting as heat carriers . following a standard calculation,@xcite @xmath18
can be obtained from the kinetic formula @xmath21 , where @xmath22 is the phonon specific heat and @xmath23 is the averaged sound velocity , and both @xmath24 and @xmath23 are known experimentally for nco.@xcite the obtained @xmath18 is compared with the averaged sample width @xmath19 (= 0.473 mm ) , which is taken to be @xmath25 times the geometrical mean width @xmath26,@xcite and shown in the inset to fig .
it can be seen that the ratio @xmath27 is only about 0.4 at 0.3 k , which means that the microscopic scattering on phonons are still effective at this temperature region.@xcite note that if the magnons can contribute to carrying heat at sub - kelvin temperatures , the phonon conductivity must be smaller than the experimental data and the phonon mean free path is even smaller than the calculated value .
the low-@xmath7 thermal conductivity of nco is further studied in the magnetic fields applied along the @xmath16 axis , the [ 110 ] direction , and the @xmath9 axis . for a low magnetic field along the @xmath16 axis , a distinct step - like transition shows up in the @xmath20 curves below 1.5 k and shifts to lower temperature with increasing field , as indicated by the 1.5 t and 2.5 t data in fig .
this transition is not observable down to 0.3 k when the magnetic field is larger than 3 t. it is likely a field - induced af - ferromagnetic ( spin polarized ) transition of nd@xmath3 ions . in the strongest field
( 14 t ) we applied , the low-@xmath7 conductivities are increased and a nearly @xmath28 dependence is presented , which explicitly indicates that there is strong magnon - phonon scattering in the zero field and the scattering is almost smeared out in the high field .
this can also be verified from the inset to fig .
2(a ) that the phonon mean free path in the 14 t field , which is much larger than that in the zero field , increases with lowering temperature and approaches the averaged sample width at 0.3 k , demonstrating that the boundary scattering limit of phonons is nearly established in the 14 t field and at such low temperatures.@xcite for magnetic field applied along the [ 110 ] direction , a similar result to the @xmath16-axis - field case is that the 14 t field can also significantly enhance the low-@xmath7 thermal conductivity ; in particular , the thermal conductivity at the lowest temperature is almost the same as that in 14 t @xmath29 .
however , there seems no af - ferromagnetic transition of nd@xmath3 ions for @xmath30 [ 110 ] , as shown in fig .
when the @xmath9-axis field is applied , the thermal conductivities display the most complex dependence on temperature , as shown in fig .
cuo@xmath2 single crystal with the field applied along the @xmath16 axis ( a ) , the [ 110 ] direction ( c ) and the @xmath9 axis ( e ) . panels ( b ) , ( d ) and ( f ) zoom in the low - field plots of panels ( a ) , ( c ) and ( e ) , respectively.,width=321 ] the detailed dependencies of thermal conductivity on the magnetic field are shown in fig .
3 . one can see from figs .
3(e ) and 3(f ) that as the @xmath9-axis field increases , @xmath0 is decreased first and then increased at high - enough field , which results in a broad valley " in the low-@xmath7 @xmath14 isotherms .
since both the cu@xmath4 spin structure and the nd@xmath3 spin structure are not changeable for @xmath31,@xcite this observation can only be attributed to the paramagnetic moments . in this
regard , the shifting of the position of the @xmath14 minimum to higher field with increasing temperature is indeed compatible with the typical behaviors of the paramagnetic moments scattering on phonons.@xcite in nco , the free " spins at very low temperatures can be either the spin vacancies / defects on the long - range - ordered cu@xmath4 spin lattice or those on the ordered nd@xmath3 spin lattice . it is known from the earlier studies on plco and gdbaco@xmath1o@xmath32 that whether the high - field - limit conductivities are larger than the zero - field values is determined by the condition whether the magnetic ions have the zero - field energy splitting.@xcite since the low-@xmath7 @xmath0 in the high - field limit is apparently larger than those in zero field , it is likely that the free " spins on the nd@xmath3 sites , whose ground - state doublet can be split in the zero field,@xcite rather than the cu@xmath4 free spins,@xcite are responsible for scattering phonons . in addition , it is clear that the decrease of @xmath0 in the 14 t field above 5 k , in contrast to the increase below 2 k , is simply because the field is not strong enough to remove the paramagnetic scattering.@xcite note that the paramagnetic moments scattering on phonons seems to be much more significant in nco than that in plco.@xcite the phonon scattering by paramagnetic moments is known to be qualitatively isotropic for different field directions,@xcite so the similar @xmath14 behaviors to those for @xmath31 should also contribute to the @xmath14 data for @xmath33 .
in fact , at relatively high temperatures above @xmath10 2 k , the profiles of @xmath14 for @xmath34 or [ 110 ] could still be understood in the picture of paramagnetic scattering , whereas at lower temperatures ( @xmath35 1 k ) the @xmath14 isotherms display very different behaviors , which manifests that some other mechanisms are taking place .
for @xmath36 $ ] , @xmath0 shows a step - like increase at @xmath10 0.6 t and a weak field dependence above this transition field , as shown in figs .
3(c ) and 3(d ) .
the situation for @xmath11 is remarkably more complicated .
as can be seen from fig .
3(a ) , at sub - kelvin temperatures , except for a sudden increase at @xmath10 1 t there is a sharp drop at @xmath10 2.5 t. one possible mechasim for these complicated @xmath14 behaviors is naturally related to the transitions of spin structures induced by the in - plane field .
it is known that the @xmath16-axis field and the [ 110 ] field cause the spin - flop transitions of the cu@xmath4 sublattice at @xmath10 4.5 t and 0.75 t , respectively.@xcite apparently , these re - orientations have weak effects on the heat transport , since there is only a very small dip in @xmath14 curves at 4.5 t for @xmath11 , as shown in fig .
3(a ) and the effect for @xmath30 [ 110 ] is not distinguishable from the strong increase of @xmath0 at 0.6 t. therefore , the low - field anomalies of @xmath14 for @xmath11 or [ 110 ] are most likely caused by the transitions of the nd@xmath3 sublattice , which have never been explored in experiments at sub - kelvin temperatures .
cuo@xmath2 in the magnetic fields along two different in - plane directions .
( a ) spin structure in the zero field .
( b - d ) spin structures in the magnetic field along the @xmath16 axis . @xmath37 and @xmath38 are the critical fields of the spin flop and the spin polarization transitions of the nd@xmath3 spins , respectively .
( e ) spin structure in the magnetic field along the [ 110 ] direction .
@xmath39}$](cu ) and @xmath40}$](cu ) are the spin - flop transition fields of the cu@xmath4 spins for @xmath11 and [ 110 ] , respectively.,width=321 ] the above @xmath14 behaviors can suggest the evolution of nd@xmath3 spin structure ( at sub - kelvin temperatures ) for the in - plane magnetic fields , which are summarized in fig .
4 . in principle , whether the nd sublattice rotates together with the cu@xmath4 spins depends on the competition between nd@xmath3-cu@xmath4 and nd@xmath3-nd@xmath3 interactions . as already known ,
nco exhibits a noncollinear structure in the zero field , as shown in fig .
4(a ) . with increasing field along the @xmath16 axis ,
the nd@xmath3 ions are possible to make spin rotations if the field is large enough . as shown in fig .
4(b ) , when @xmath41 1 t ) those nd@xmath3 spins pointing along the field direction could rotate by 90@xmath42 in the @xmath15 plane , in other words , a spin flop transition of nd@xmath3 ions happens .
it is known that the low - energy magnons are usually well populated at the spin flop transition due to the close of the magnon gap,@xcite and meanwhile @xmath0 shows a sudden increase at the transition field .
it is therefore natural to conclude that the nd@xmath3 magnons act as heat carriers at very low temperatures and hence make a positive contribution to the thermal conductivity . if the magnetic field increases further , the nd@xmath3 spins gradually turn to the direction of applied field and are finally polarized at @xmath43 ( @xmath10 2.5 t ) , as illustrated in fig .
4(c ) , which corresponds to a dip in @xmath14.@xcite after that the thermal conductivity starts to recover due to the weakening of magnon scattering on phonons . at @xmath10 4.5 t , the cu@xmath4 spin flop happens and the spin directions are switched to be perpendicular to the field .
this field strength can destroy the intra - unit nd@xmath3-cu@xmath4 interaction , therefore nd@xmath3 spins can no longer rotate together with cu@xmath4 spins and keep the alignment along the @xmath16 axis , as shown in fig .
note that in the high - field limit , the roles of magnons as either heat carriers or phonon scatterers are inactive and the plateau of @xmath14 also indicates that the paramagnetic scattering on phonons is smeared out , as the 14 t @xmath20 data indicate . in the case of @xmath12 $ ] ,
there is also a sharp increase of @xmath0 at @xmath10 0.6 t and the magnitude of increase is almost the same as the increase at 1 t @xmath29 .
this strongly suggests a magnon heat - transport contribution appearing at this critical field , which is essentially the same as that for the cu@xmath4 spin reorientation ( @xmath10 0.75 t in some former reports).@xcite since this field strength is much smaller than the nd@xmath3-cu@xmath4 interaction , it is expectable that the nd@xmath3 spins rotate together with the cu@xmath4 spins and thereby align along the field direction , as shown in fig .
similarly , the low - energy nd@xmath3 magnons are well populated at this spin flop transition and are able to significantly transport heat .
it is notable that there is no dip - like feature in higher fields , which indicates that the nd@xmath3 spin polarization does not happen .
this is supported by the @xmath20 in fig .
2(b ) , where there is no af - ferromagnetic transition behavior similar to those in fig .
it is useful to make a quantitative estimation on the nd@xmath3 magnon heat transport from the above data .
for example , the step - like increases of @xmath0 at 0.36 k is about 0.7 times the zero - field value , which gives a magnon thermal conductivity of 0.0441 w / km , for both @xmath11 and @xmath44 [ 110 ] . taking a theoretical prediction of the velocity of nd@xmath3 magnons , @xmath10 10 mev(no experimental observation so far),@xcite and assuming the ballistic transport of magnons at such low temperature,@xcite
we can obtain the magnon mean free path of about 0.16 mm .
this value is quite reasonable since it is in the same order of magnitude of the averaged sample width ( 0.473 mm ) . with above understandings on both the paramagnetic scattering and the magnon heat transport associated with spin - structure transitions
, one can get a complete picture of the @xmath14 isotherms for the in - plane fields . at sub - kelvin temperatures , the sharp increase and the dip " at low fields along the @xmath16 axis
are mainly related to the magnon behaviors , while the high - field enhancement and the plateau feature are due to the disappearance of paramagnetic scattering . on the other hand , the step increase at low fields along [ 110 ]
mainly results from the magnon heat transport , and the gradual increase of @xmath0 at higher field up to @xmath10 6 t is likely due to the weakening of paramagnetic scattering ; furthermore , the slow decrease of @xmath0 at field above 6 t is related to the suppression of low - energy magnon excitations in very high fields . in passing ,
it is worthy of pointing out that at 1@xmath102 k the thermal conductivity is most strongly recovered at very high in - plane field .
it seems to be related to the temperature dependence of magnon excitations ( nd@xmath3 ) ; that is , at @xmath35 1 k , the magnon population is negligibly small in zero field because of the magnon gap , while at 1@xmath102 k , the critical region of nd@xmath3 af transition , the spin fluctuations are significant and scatter phonons rather strongly .
so in this temperature region the zero - field phonon transport is most strongly damped by not only the paramagnetic moments but also the magnetic excitations and it can be remarkably recovered by applying high in - plane field .
it is intriguing to compare the present work with the earlier ones .
it has been concluded from the very - low-@xmath7 thermal conductivity in the zero field and strong in - plane field , which is much higher than the cu@xmath4 spin - flop transition , that the high - field - induced enhancement of @xmath0 is a direct contribution from nd@xmath3 magnon heat transport.@xcite the present data essentially support the capability of nd@xmath3 magnon transporting heat , but only in relatively low fields .
base on the detailed field dependence of @xmath0 , particularly for @xmath45 , it is able to be clarified that the enhancement of @xmath0 at very high fields is mainly due to the weakening of paramagnetic scattering on phonons .
the reason that the earlier work did not notice the importance of paramagnetic scattering is that the @xmath20 data for @xmath31 were taken only at 10 t for the comparison with the in - plane - field data.@xcite it is a coincidence that in this field , the thermal conductivities in the whole temperature range are not likely to be larger than the zero - field values , as shown in fig .
therefore , the present data provide a very important supplement to the earlier works and lead to a more accurate conclusion .
in summary , we study the heat transport of a parent insulator of high-@xmath46 superconductors nd@xmath1cuo@xmath2 at low temperatures down to 0.3 k and in magnetic fields up to 14 t. it is found that in zero field the low-@xmath7 thermal conductivity is purely phononic with rather strong scatterings from paramagnetic moments and nd@xmath3 magnetic excitations . in high magnetic field along either the @xmath9 axis or the @xmath15 plane , the low-@xmath7 thermal conductivity can be significantly enhanced because of the weakening of magnetic scattering .
an interesting finding is the drastic changes of @xmath0 at low fields along the @xmath16 axis or the [ 110 ] direction , which demonstrates the field - induced spin flop or spin polarization of nd@xmath3 spin lattice . at sub - kelvin temperatures
, the nd@xmath3 magnons can act as heat carriers in the spin - flopped state , however , their transport can exist only in some intermediate field regime and is suppressed by an succeeding spin - polarization transition for @xmath11 . in the field along the [ 110 ] direction , the magnon transport can be active in a broader field region but is also weakened in very high field .
an evolution of the magnetic structure with the in - plane field can be suggested based on these field dependencies of @xmath0 .
one may note that although the magnons can hardly affect the phonon heat transport for those superconducting samples , because of the disappearance of the long - range af order , the paramagnetic moments can still effectively scatter phonons .
this work was supported by the chinese academy of sciences , the national natural science foundation of china , the national basic research program of china ( grants no .
2011cba00111 and no .
2009cb929502 ) , and the rfdp ( grant no .
20070358076 ) .
m. f. hundley , j. d. thompson , s .- w .
cheong , z. fisk , and s. b. oseroff , physica c * 158 * , 102 ( 1989 ) ; n. t. hien , v. h. m. duijn , j. h. p. colpa , j. j. m. franse , and a. a. menovsky , phys .
b * 57 * , 5906 ( 1998 ) ; r. kuentzler , g. pourroy , a. tigheza , and y. dossmann , solid state commun . *
78 * , 113 ( 1991 ) . | we report the magnetic - field dependence of thermal conductivity ( @xmath0 ) of an insulating cuprate nd@xmath1cuo@xmath2 at very low temperatures down to 0.3 k. it is found that apart from the paramagnetic moments scattering on phonons , the nd@xmath3 magnons can act as either heat carriers or phonon scatterers , which strongly depends on the long - range antiferromagnetic transition and the field - induced transitions of spin structure . in particular , the nd@xmath3 magnons can effectively transport heat in the spin - flopped state of the nd@xmath3 sublattice .
however , both the magnon transport and the magnetic scattering are quenched at very high fields .
the spin re - orientations under the in - plane field can be conjectured from the detailed field dependence of @xmath0 . |
the present paper , devoted to the scalar wave equation in a non - commutative schwarzschild space - time , is strongly motivated by three branches of modern gravitational physics : 0.3 cm ( i ) in their investigation of quantum amplitudes in black - hole evaporation @xcite , the authors of @xcite have considered emission of scalar radiation in a black - hole collapse problem , assuming non - spherical perturbations of the scalar field @xmath2 on the final surface @xmath3 , and that the intrinsic three - metric describes an exactly spherically - symmetric spatial gravitational field . 0.3 cm ( ii ) in general relativity , unexpected features of the asymptotic structure
are already found to occur : massless scalar fields which have a bondi - type expansion in powers of @xmath4 near null past infinity do not have such an expansion near future null infinity ; solutions which have physically reasonable cauchy data may fail to have bondi - type expansions near null infinity @xcite .
0.3 cm ( iii ) according to the models studied in @xcite , @xcite , @xcite , the non - commutativity of spacetime can be encoded in the commutator of operators corresponding to spacetime coordinates , i.e. ( the integer @xmath5 below being even ) @xmath6={\rm i } \ ; \theta^{\mu \nu } , \ ; \mu,\nu=1,2, ... ,d \label{(1.1)}\ ] ] when the antisymmetric matrix @xmath7 is taken to have a block - diagonal form @xmath8 with @xmath9 the parameter @xmath1 having dimension of length squared and being constant .
as shown in @xcite , the constancy of @xmath1 is very important to obtain a consistent treatment of lorentz invariance and unitarity .
the authors of @xcite solve the einstein equations with mass density of a static , spherically symmetric , smeared particle - like gravitational source as ( hereafter , in agreement with our earlier work @xcite , we use @xmath10 units ) @xmath11 which therefore plays the role of matter source .
their resulting spherically symmetric metric is @xmath12 where , in terms of the lower incomplete gamma function @xmath13 we define the mass function @xcite @xmath14 thus , if one tries to study emission of scalar radiation as in @xcite but in the presence of a non - vanishing @xmath1 parameter ( cf @xcite for the pure gravity case ) , one is naturally led to study a scalar wave equation in a spherically symmetric spacetime whose metric is affected by @xmath1 , which is the goal of the present paper .
section 2 builds conformal infinity for the space - time with metric ( 1.4 ) .
section 3 , following @xcite , turns the scalar wave equation into an inhomogeneous euler poisson darboux equation .
section 4 solves such an equation and shows under which conditions there is full qualitative agreement with general relativity .
concluding remarks and open problems are presented in section 5 , while the appendices describe relevant mathematical details .
inspired by general relativity , we define a new radial coordinate @xmath15 in such a way that @xmath16 this equation is solved by @xmath17 and the metric ( 1.4 ) can be re - expressed in the form @xmath18 where @xmath19 , \varphi \in [ 0,2\pi]$ ] , and we have defined the ` retarded ' coordinate @xmath20-\infty,\infty [ , \label{(2.4)}\ ] ] and the ` advanced ' coordinate @xmath21-\infty,\infty[. \label{(2.5)}\ ] ] the equations ( 2.2 ) , ( 2.4 ) and ( 2.5 ) yield @xmath22}. \label { ( 2.6)}\ ] ] to lowest order , eq .
( 2.6 ) is solved by @xmath23 , and on defining @xmath24 one finds , by iterated approximations , the asymptotic expansion @xmath25 the limit @xmath26 with @xmath27 fixed defines future null infinity @xmath28 ; the limit @xmath29 with @xmath30 fixed defines past null infinity @xmath31 , while the limit @xmath32 with @xmath33 fixed defines spacelike infinity , i.e. the point @xmath34 .
the figures below show the behaviour of the denominator @xmath35 in eq .
( 2.1 ) for various values of @xmath1 .
interestingly , the occurrence of @xmath1 does not introduce new singularities with respect to general relativity .
it is actually simpler to introduce the coordinates @xmath36 and @xmath37 separately , which yields the conformally rescaled , `` unphysical '' metrics ( here @xmath38 , and @xmath39 is the metric on a unit two - sphere ) @xmath40 = -(f^{2}-2mf^{3})du^{2}+2du \ ; df + d\sigma^{2 } , \label{(2.9)}\ ] ] and @xmath41 these metrics are manifestly regular and analytic on their respective hypersurfaces @xmath42 , since their determinants are equal to @xmath43 for all @xmath44 , including @xmath42 .
the physical space - time corresponds to @xmath45 in ( 2.9 ) , and we can extend the manifold to include @xmath28 , given when @xmath42 . similarly , in ( 2.10 ) , the physical space - time corresponds to @xmath45 and can be extended to include @xmath31 , given when @xmath42 .
only the boundary @xmath46 is adjoined to the space - time . in common with general relativity ,
we note here a difficulty that is encountered if we try to identify @xmath31 with @xmath28 .
if we do extend the region of definition of ( 2.9 ) to include negative values of @xmath44 , and then make the replacement @xmath47 , we see that the metric has the form ( 2.10 ) ( with @xmath36 in place of @xmath37 ) but with a mass function @xmath48 in place of @xmath49 .
thus , the extension across @xmath50 involves a reversal of the sign of the mass function , which is incompatible with eq .
( 1.6 ) unless we advocate a discontinuity in the derivative of the curvature @xcite across @xmath51 .
it is therefore not reasonable to identify @xmath28 with @xmath31 . to sum up
, we have two disjoint boundary hypersurfaces @xmath28 and @xmath31 , each of which is a cylinder with topology @xmath52 .
it is clear from ( 2.9 ) and ( 2.10 ) that each of @xmath53 is a null hypersurface ( the induced metric at @xmath42 being degenerate ) .
these null hypersurfaces are generated by rays ( given by @xmath54 , @xmath42 ) whose tangents are normals to the hypersurfaces .
these rays may be taken to be the @xmath55 s of the topological product @xmath52 .
the coordinates @xmath56 defined in ( 2.4 ) and ( 2.5 ) are not the most convenient for discussing the limits which define conformal infinity @xcite .
we therefore define ( cf . @xcite ) a function @xmath57 by requiring that @xmath58 should be equal to @xmath15 in ( 2.2 ) , i.e. @xmath59 which implies @xmath60 equation ( 3.1 ) defines a one - parameter family of monotone decreasing @xmath61 functions taking values over the whole real line .
the monotone decreasing character of @xmath62 is proved by imposing that @xmath63 .
this is indeed satisfied for sufficiently small values of @xmath1 , so that @xmath64 , which is positive provided @xmath65 .
a @xmath61 inverse function therefore exists , which makes it possible to define new coordinates @xmath66 according to ( cf @xcite ) @xmath67 @xmath68 where the integrals ( 3.3 ) and ( 3.4 ) involve the mass function @xmath69 . on defining @xmath38 as in section 2
, one finds from ( 2.2 ) , ( 3.1 ) , ( 3.3 ) and ( 3.4 ) that @xmath70 moreover , from ( 3.2)(3.4 ) , the metric ( 2.3 ) in the @xmath71 coordinates takes the following form in the @xmath72 coordinates : @xmath73 having defined @xmath74 @xmath75 in the analysis of the scalar wave equation @xmath76 , we now rescale the scalar field @xmath77 according to @xmath78 where @xmath79 is a real positive function such that @xmath80 the ` unphysical ' scalar field @xmath81 satisfies the conformally invariant wave equation in @xmath82 spacetime dimensions , i.e. @xmath83 where @xmath84 is the dalembert wave operator @xcite , and @xmath85 is the scalar curvature in the ` unphysical ' space - time with line element @xmath86 on choosing the conformal factor in the form @xmath87 , we therefore obtain the metric tensor @xmath88 having introduced @xmath89 @xmath90 @xmath91 by virtue of spherical symmetry , we look for solutions of eq .
( 3.11 ) as a linear combination of factorized terms as @xmath92 where @xmath93 are the spherical harmonics on @xmath94 .
substitution of the ansatz ( 3.17 ) into eq .
( 3.11 ) gives @xmath95=s_{\theta}(a , b)\chi , \label{(3.18)}\ ] ] where @xmath96 is the euler poisson darboux operator @xcite @xmath97 which depends implicitly on @xmath1 through @xmath66 defined in ( 3.3 ) , ( 3.4 ) , while @xmath98 is the @xmath1-dependent source term @xmath99 \nonumber \\ & + & { 1\over 6}(a+b)^{2 } g_{1}(a)g_{2}(b)f(a , b)r(\theta ) , \label{(3.20)}\end{aligned}\ ] ] having denoted by @xmath100 the scalar curvature in eq .
inspired by @xcite , we now write the solution of eq .
( 3.18 ) as the sum @xmath101 , where @xmath102 is the general solution of the homogeneous euler poisson darboux equation @xmath103=0 $ ] , while @xmath104 is an integral operator with kernel given by the riemann green function ( see appendix ) of @xmath96 @xcite : @xmath105 having defined @xmath106 as is described in @xcite , @xmath107 has the general form @xmath108 , \label{(3.23)}\ ] ] with @xmath109 and @xmath110 arbitrary @xmath111 functions .
moreover , the riemann green function of the operator @xmath96 defined in ( 3.19 ) is obtained from the legendre polynomial of degree @xmath112 according to @xcite @xmath113 having defined @xcite @xmath114
hereafter we consider for simplicity the case @xmath115 ; the comparison with our ref .
@xcite is then easier , and all main features are already displayed .
strictly , we consider an asymptotic characteristic initial - value problem where data are specified on past null infinity for @xmath116 $ ] and on the outgoing null hypersurface @xmath117 . if @xmath115 , it is clear from ( 3.23 ) that the characteristic data can be set to @xmath118 : @xmath119 , while the riemann green function in ( 3.24 ) reduces to @xmath118 @xcite : @xmath120 the inhomogeneous wave equation ( 3.18 ) with @xmath115 can be solved with the help of a contraction mapping , i.e. @xcite @xmath121 where @xmath122 and the series in ( 4.2 ) is known to be uniformly convergent near spacelike infinity in general relativity @xcite . moreover , in general relativity the partial derivative @xmath123 as @xmath124 and @xmath125 is fixed remains bounded , as well as the partial derivative @xmath126 as @xmath127 and @xmath128 is fixed .
the second derivative @xmath129 , however , diverges near future null infinity , which implies that the presence of a conformal singularity at spacelike infinity affects the behaviour of scalar fields on future null infinity , reducing by a considerable amount their differentiability class .
such a property is proved by exploiting the integral representation of @xmath129 , i.e. @xmath130db ' . \label{(4.4)}\ ] ] by insertion of ( 4.2 ) into ( 4.4 ) , and bearing in mind ( 4.3 ) , one finds that the possible singularities of @xmath129 are ruled by the integrals @xcite @xmath131 @xmath132 where @xmath133 remains finite as @xmath124 , whereas @xmath134 displays a logarithmic singularity as @xmath124 .
if @xmath115 in sec .
3 we find , for finite values of @xmath1 , the counterpart of ( 4.5 ) given by the integral @xmath135 where , from the asymptotic formulae as @xmath124 and @xmath125 is fixed , we find , for all finite values of @xmath1 , @xmath136 @xmath137 @xmath138 @xmath139 @xmath140 , \label{(4.12)}\ ] ] and hence @xmath141 in agreement with the analysis in @xcite for general relativity .
these approximations should be abandoned only if @xmath1 is so large that ( cf .
( 4.9 ) ) @xmath142 furthermore , the counterpart of ( 4.6 ) is given by the integral @xmath143 again in full agreement with @xcite .
note that a more accurate asymptotic expansion of the source term would be @xmath144 but this does not modify the leading terms as @xmath124 in ( 4.13 ) and ( 4.15 ) .
the figures below show the behaviour of @xmath145 and @xmath129 .
ever since penrose @xcite developed a geometrical picture of an isolated system in general relativity as a space - time admitting future and past null infinity ( with the associated fall - off of the metric along null geodesics going off to infinity ) , there has been always great interest in this coordinate - free way of bringing infinity to a ` finite distance ' and discussing the asymptotic structure of space - time .
moreover , the conceptual revolution brought about by non - commutative geometry @xcite has led to an assessment of the very concept of space - time manifold @xcite , with ` corrections ' to it evaluated , for example , along the lines of the work in refs .
@xcite . within this broad framework ,
the contributions of our paper are as follows . 0.3 cm ( i ) construction of conformal infinity for the spherically symmetric space - time which incorporates noncommutative - geometry corrections to schwarzschild space - time
. 0.3 cm ( ii ) evaluation of the source term ( 3.20 ) in the inhomogeneous euler poisson darboux equation which describes the scalar wave equation in the unphysical space - time obtained after conformal rescaling of the original metric ( 1.4 )
. 0.3 cm ( iii ) qualitative analysis of the asymptotic characteristic initial - value problem in the @xmath115 case , finding again the logarithmic singularity as shown in eq .
( 4.15 ) . in the original , ` physical ' space - time with metric ( 1.4 ) , such a singularity implies that the large-@xmath0 behaviour of the scalar field is described by the asymptotic expansion @xcite @xmath146 and therefore the field falls off at large @xmath0 rather more slowly than in flat space - time @xcite .
0.3 cm ( iv ) numerical support for all results in sections 3 and 4 , as shown by figures 57 at the end of section 4 .
our results are thus an encouraging progress towards a rigorous theory of wavelike phenomena in noncommutative geometry , along the lines of the conformal - infinity program of penrose for general relativity @xcite .
hopefully , the physical applications to isolated gravitating systems in a noncommutative framework , and possibly to black - hole evaporation , will also become clear in the years to come .
the authors are grateful to the infn for financial support .
the work of g. miele has been partially supported by prin _ fisica astroparticellare_.
an hyperbolic equation in two independent variables can be always cast in the canonical form @xcite
@xmath95=\left({\partial^{2}\over \partial x \partial y } + a(x , y){\partial \over \partial x}+b(x , y){\partial \over \partial y } + c(x , y ) \right)\chi(x , y)=h(x , y ) .
\label{(a1)}\ ] ] one can then use the riemann integral representation of the solution @xcite . for this purpose ,
denoting by @xmath147 the adjoint of the operator @xmath96 in ( a1 ) , which acts according to @xmath148=\chi_{,xy}-(a \chi)_{,x } -(b \chi)_{,y}+c \chi , \label{(a2)}\ ] ] one has to find the riemann kernel @xmath149 subject to the following conditions ( @xmath150 being the coordinates of a point @xmath151 such that the characteristics through it intersect a curve @xmath152 at points @xmath109 and @xmath110 , @xmath153 being a segment with constant @xmath154 , and @xmath155 being a segment with constant @xmath156 ) : 0.3 cm ( i ) as a function of @xmath156 and @xmath154 , @xmath85 satisfies the adjoint equation @xmath157=0 , \label{(a3)}\ ] ] 0.3 cm ( ii ) @xmath158 on @xmath153 , i.e. @xmath159 and @xmath160 on @xmath155 , i.e. @xmath161 0.3 cm ( iii ) @xmath85 equals @xmath118 at @xmath151 , i.e. @xmath162 it is then possible to express the solution of eq .
( a1 ) in the form @xcite @xmath163 + \int_{ab } \biggr ( \left[{r\over 2}\chi_{,x } + \left(br -{1\over 2}r_{,x}\right)\chi \right]dx \nonumber \\ & - & \left[{r\over 2}\chi_{,y}+\left(ar -{1\over 2}r_{,y } \right)\chi \right]dy \biggr ) + \int \int_{\omega}r(x , y;\xi,\eta)h(x , y)dxdy , \label{(a7)}\end{aligned}\ ] ] where @xmath79 is a domain with boundary .
note that eqs .
( a4 ) and ( a5 ) are ordinary differential equations for the riemann function @xmath149 along the characteristics parallel to the coordinate axes . by virtue of ( a6 ) , their integration yields @xmath164 @xmath165 which are the values of the riemann kernel @xmath85 along the characteristics through @xmath151 . equation ( a7 ) yields instead the solution of eq .
( a1 ) for arbitrary initial values given along an arbitrary non - characteristic curve @xmath152 , by means of a solution @xmath85 of the adjoint equation ( a3 ) which depends on @xmath166 and two parameters @xmath167 . unlike @xmath168 ,
the riemann function solves a characteristic initial - value problem @xcite .
in section 3 , the evaluation of the scalar curvature @xmath169 in the unphysical space - time with metric ( 3.13 ) is as follows . with the notation in eqs .
( 3.13)(3.16 ) , the non - vanishing connection coefficients turn out to be ( no summation over repeated indices ) @xmath170 @xmath171 @xmath172 @xmath173 @xmath174 @xmath175 the resulting riemann tensor is evaluated from the general formula in a coordinate basis @xmath176 since we evaluate the scalar curvature @xmath177 , which in our case is equal to ( hereafter , no summation over repeated indices @xmath128 or @xmath125 or @xmath178 or @xmath179 ) @xmath180 we only need the @xmath181 components of riemann occurring in @xmath182 @xmath183 @xmath184 among these , the only non - vanishing are @xmath185 @xmath186 @xmath187 and hence @xmath188 with the notation in eqs .
( 3.14)(3.16 ) , we find @xmath189 , \label{(b16)}\ ] ] @xmath190 , \label{(b17)}\ ] ] @xmath191 .
\label{(b18)}\ ] ] therefore eq .
( b15 ) yields @xmath192 . \label{(b19)}\ ] ] for example , in general relativity @xmath193 by virtue of @xmath194 and hence @xmath195 @xmath196 our definition of scalar curvature has opposite sign with respect to the work in @xcite , but of course this does not affect the results . hawking s w 2005 _ phys .
_ d * 72 * 084013 farley a n st j and death p d 2004 _ phys .
_ b * 601 * 184 farley a n st j and death p d 2005 ` quantum amplitudes in black - hole evaporation .
spin-0 amplitude ' ( gr - qc/0510029 ) schmidt b g and stewart j m 1979 _ proc .
roy . soc .
a * 367 * 503 nicolini p , smailagic a , and spallucci e 2006 _ phys .
_ b * 632 * 547 smailagic a and spallucci 2004 _ j. phys . _ a * 37 * 7169 spallucci e , smailagic a and nicolini p 2006 _ phys .
_ d * 73 * 084004 di grezia e , esposito g , and miele g 2006 _ class .
quantum grav . _
* 23 * 6425 penrose r and rindler w 1986 _ spinors and space - time .
2 . spinor and twistor methods in space - time geometry _
( cambridge : cambridge university press ) courant r and hilbert d 1961 _ methods of mathematical physics .
ii . partial differential equations _ ( new york : interscience ) copson e t 1958 _ arch . ration .
mech . analysis _
* 1 * 324 penrose r 1964 ` conformal treatment of infinity ' in _ relativity , groups and topology _ , eds .
c. dewitt and b. dewitt ( new york : gordon & breach ) landi g 1997 ` an introduction to noncommutative spaces and their geometry ' ( hep - th/9701078 ) connes a 2006 _ aip conf .
proc . _ * 861 * 47 ; connes a 2000 _ j. math .
phys . _ * 41 * 3832 lizzi f 2007 _ int . j. mod
. phys . _ a * 22 * 1317 aschieri p , dimitrijevic m , meyer f and wess j 2006 _ class .
quantum grav . _
* 23 * 1883 | recent work in the literature has studied a version of non - commutative schwarzschild black holes where the effects of non - commutativity are described by a mass function depending on both the radial variable @xmath0 and a non - commutativity parameter @xmath1 .
the present paper studies the asymptotic behaviour of solutions of the zero - rest - mass scalar wave equation in such a modified schwarzschild space - time in a neighbourhood of spatial infinity .
the analysis is eventually reduced to finding solutions of an inhomogeneous euler poisson darboux equation , where the parameter @xmath1 affects explicitly the functional form of the source term .
interestingly , for finite values of @xmath1 , there is full qualitative agreement with general relativity : the conformal singularity at spacelike infinity reduces in a considerable way the differentiability class of scalar fields at future null infinity . in the physical space - time
, this means that the scalar field has an asymptotic behaviour with a fall - off going on rather more slowly than in flat space - time .
the scalar wave equation in a non - commutative spherically symmetric space - time |
we consider lepton - mass effects in the off - shell decays of gauge bosons in the processes @xmath25 and @xmath26 where the off - shell gauge bosons @xmath27 are produced in the higgs decays @xmath10 , @xmath28 . in the @xmath10 case
the corresponding @xmath29 modes have recently been observed at the lhc and are therefore adequately dubbed `` higgs discovery channels '' @xcite .
further evidence on these decays has been presented in ref .
the quantum numbers of the higgs boson have been pinned down by an angular analysis of the four leptons in the final state to be @xmath30 both in the leptonic @xmath10 mode @xcite as well as in the leptonic @xmath31 mode @xcite . on the theoretical side
there have been a number of papers analyzing the quantum numbers of the higgs boson through an angular analysis of the four - lepton final state among which are refs .
the physics of the higgs boson in all its aspects has been nicely reviewed in three recent papers @xcite .
off - shell effects in the decays involving massive leptons will lead to additional scalar and scalar longitudinal interference contributions well familiar from neutron beta decay , the semileptonic decay @xmath32
@xcite , or from the decays @xmath33 @xcite and @xmath34 @xcite .
the scalar and scalar longitudinal interference contributions are quadratic in the lepton masses and can thus be neglected at the scale @xmath5 . however , for the off - shell decays @xmath35 the scale is not set by @xmath5 but by the off - shellness of the respective gauge bosons which extends from threshold @xmath36 ( maximal recoil point ) to the zero recoil point at @xmath37 , i.e. one has @xmath38 one will therefore have to carefully consider @xmath0-lepton mass effects particularly in the @xmath3 region close to threshold given by @xmath39 and @xmath40 for the leptonic modes in the decays @xmath10 and @xmath41 , respectively .
lepton - mass effects reduce the overall rate relative to the zero lepton - mass case .
in addition , lepton - mass effects lead to leptonic helicity - flip contributions which in turn can generate novel angular dependencies in the respective angular decay distributions
. these angular dependencies can mimic new angular terms introduced by higher dimension effective coupling terms @xcite or non - sm @xmath42 coupling terms @xcite .
@xmath0-lepton mass effects should therefore not be neglected if one is aiming for high precision physics in the higgs sector . and @xmath43
are also taken into account in ref .
@xcite . ]
our paper is structured as follows .
after this introductory section , in sec
. 2 we present a general formula for the three - fold angular angular decay distribution for the on - shell off - shell decays @xmath44 .
the angular decay distribution is obtained using helicity methods . in sec .
3 we discuss lepton - mass effects in the decay @xmath45 and their effect on the rates and the angular decay distributions .
we do the same in sec . 4 for the decays @xmath46 .
in sec . 5
we summarize our results and conclude with some general remarks . some technical material regarding helicity amplitudes
is relegated to the appendices . in appendix
a we list the helicity amplitudes for the @xmath47 transitions .
the helicity representation of the lepton tensors in the neutral- and charged - current cases can be found in appendix b and c , respectively .
the three - fold angular decay distribution in the cascade decays @xmath48 can be derived from the covariant contraction of the on - shell and off - shell lepton tensors @xmath49 and @xmath50 with the @xmath42 higgs coupling @xmath51 where the vertices are connected by the propagator projectors @xmath52 ( spin 1 ) and @xmath53 ( spin @xmath54 ) .
one has @xmath55 where , in the standard model ( sm ) , @xmath56 .
we denote the on - shell and off - shell momenta of the gauge bosons by @xmath57 and @xmath58 . in the unitary gauge
the on - shell spin-1 propagator @xmath52 ( @xmath59 ) and the off - shell propagator @xmath53 ( @xmath60 ) read @xmath61 note that in the unitary gauge gauge where one has to consider goldstone boson exchange in addition to gauge boson exchange . in the coupling to the final state fermion pair the gauge parameter @xmath62 cancels between the goldstone and gauge boson contributions , resulting in the unitary gauge propagator .
this has been explicitly demonstrated for fermion fermion scattering @xcite and for the decay @xmath63 @xcite . ]
the off - shell propagator @xmath64 contains a spin-1 and a spin-0 piece .
this can be seen by splitting the off - shell gauge propagator in eq .
( [ offshell ] ) into its spin-1 and spin-0 components according to @xmath65 where @xmath66 in the zero lepton - mass approximation one has @xmath67 and therefore the spin-0 piece in eq .
( [ split ] ) does not contribute and can be dropped when evaluating eq .
( [ angdis1 ] ) .
this is always a good approximation for @xmath29 but no longer a good approximation for @xmath68 .
an interesting observation concerns the spin-0 contribution . taken together with the propagator pole proportional to @xmath69
, the contribution of the spin-0 piece can be seen to be proportional to a contact interaction of the form @xmath70 with a @xmath71-dependent coupling when one sets @xmath72 .
technically there are two routes to obtain angular decay distributions from eq .
( [ angdis1 ] ) . in the first route one
parametrizes the four - vectors of the problem in terms of the five phase - space variables @xmath59 , @xmath3 , @xmath73 , @xmath74 and @xmath75 ( cf.figs . [ planeszz ] and [ planesww ] ) .
the covariant evaluation of the lorentz - invariant expression ( [ angdis1 ] ) leads to a number of scalar products of momenta that are defined in different reference frames .
when doing the requisite contractions , the four - momenta have to be boosted to a common reference frame as e.g. described in ref .
@xcite for the decay @xmath76 and in ref .
@xcite for the decay @xmath77 .
one then arrives at the desired three - fold joint angular decay distribution . a second ,
perhaps more intelligent route , is to use an analysis in terms of helicity amplitudes .
the advantage of the helicity method is that the origin of the angular factors multiplying the helicity structure functions can be straightforwardly identified .
the angular factors can be seen to arise from the transformation properties of the helicity amplitudes under the action of the rotation group . in order to transform to the helicity representation of the covariant form in eq .
( [ angdis1 ] ) one makes use of the completeness relation for the spin-1 on - shell and off - shell polarization vectors . the on - shell and off - shell propagator can be expanded according to @xcite @xmath78 ( @xmath59 ) and @xmath79 note that there is an additional spin-0 degree of freedom propagating in the off - shell propagator in eq .
( [ complete2 ] ) .
we shall specify this spin-0 degree of freedom by assigning the label @xmath80 ( @xmath81 for time - component ) to this mode . according to the separation in eq .
( [ split ] ) the `` @xmath81 '' mode carries the weight @xmath82 which finally leads to @xmath83 in eq .
( [ complete2 ] ) .
the four polarization four - vectors @xmath84 will be specified in appendix a. in low - energy calculations such as neutron @xmath85 decay or in the semileptonic
bottom - hadron decays one usually drops the term proportional to @xmath86 in eq .
( [ scalar ] ) since one has @xmath87 .
however , in the present application the factor @xmath88 can become as large as @xmath89 at the zero recoil point and can therefore not be neglected .
where one is aiming for ultrahigh precision , the importance of the @xmath90 contributions have been discussed in the literature @xcite . ] with the help of the completeness relations ( [ complete1 ] ) and ( [ complete2 ] ) the covariant form of the angular decay distribution ( [ angdis1 ] ) can be cast into a representation in terms of helicity components .
one has @xmath91 where @xmath92 for @xmath93 , @xmath94 for @xmath80 and correspondingly for the primed quantities .
it turns out that @xmath95 in the decay @xmath96 as long as one is not analyzing @xmath0-polarization effects , i.e. there are no spin-0
spin-1 interference effects in this decay .
the evaluation of the helicity components of the @xmath47 transition amplitudes @xmath97 is given in appendix a while the evaluation of the helicity components of the lepton tensors @xmath98 and @xmath99 are given in appendix b ( neutral - current case ) and c ( charged - current case ) . up to this point
we have allowed for a general structure of the @xmath42 coupling . in the following
we shall specify to the sm coupling with @xmath56 .
in this section we write down the three - fold angular decay distribution of the decay @xmath101 involving two different pairs of leptons , i.e. we assume @xmath102 .
the corresponding decay @xmath103 involving two pairs of identical leptons ( with and without lepton - mass effects ) is more difficult to analyze due to the presence of nonfactorizing interference contributions .
these identical - particle effects will be treated in a separate paper @xcite .
we begin our discussion by presenting an explicit form of the three - fold angular decay distribution given by eq .
( [ angdis2 ] ) .
the relevant helicity components of the on - shell and off - shell lepton tensors are listed in appendix b while the helicity components of the @xmath105 transition amplitude can be found in appendix a. the polar angles @xmath106 and @xmath107 are defined in the respective lepton pair center - of - mass systems as shown in fig .
[ planeszz ] .
the azimuthal angle @xmath75 describes the relative orientation of the two decay planes .
we split the decay distribution into a helicity - nonflip and helicity - flip part , @xmath108 and @xmath109 where @xmath110 and @xmath111 . for symmetry reasons and for later applications in the off - shell off - shell case
we have written @xmath112 for @xmath113 and @xmath110 for @xmath114 on the on - shell side .
the double spin - density matrix elements @xmath115 are bilinear forms of the helicity amplitudes describing the @xmath116 transitions .
they are defined in appendix a. in eqs .
( [ zjoint3 ] ) and ( [ zjoint4 ] ) we have also included the contributions from the parity - violating terms proportional to @xmath117 and @xmath118 .
these coefficient functions are not populated by the parity - conserving sm @xmath42 coupling . in appendix
a we briefly discuss the contribution of a parity - violating non - sm coupling proportional to @xmath119 which would populate the @xmath117 and @xmath118 coefficient functions @xcite .
we add the flip and non - flip contributions and expand the result in terms of the legendre polynomials @xmath120 and @xmath121 .
the result is written in the form @xmath122 where in the second equation of ( [ eight ] ) we have split the coefficient function @xmath123 into its helicity - flip and helicity - nonflip part using the notation @xmath124 .
the coefficient functions @xmath125 and @xmath126 and their associated angular factors @xmath127 are listed in table [ tseven ] where we use the abbreviation @xmath128 with @xmath129 .
in addition we use a short - hand notation for the double density matrix elements , namely @xmath130 , @xmath131 , @xmath132 and @xmath133 . note that we have dropped the contributions of the parity - violating terms proportional to @xmath117 and @xmath118 in table [ tseven ] which are not populated by the parity - conserving sm @xmath42 coupling . the dominant flip contributions proportional to @xmath134 are contained in @xmath135 and @xmath136 .
compared to @xmath137 the leptonic vector coupling @xmath138 is much suppressed .
this is different in the quark
antiquark case treated in refs .
@xcite where the electroweak vector and axial vector couplings to the quark pairs are comparable in size . as a result
the pattern of the helicity - flip contributions in the quark pair production case is quite different from the lepton - pair production case @xcite .
.[tseven]coefficient functions appearing in the three - fold angular decay distribution of the decay @xmath139 [ cols= " < , < , < , < " , ] we define a normalized decay distribution @xmath140 where @xmath141 ( and @xmath142 , @xmath143 ) and where @xmath144 the differential decay rate distribution is given by @xmath145 where @xmath146 ( note the additional factor @xmath147 in the numerator ) and @xmath148 . the decay rate for @xmath149 ( @xmath150 )
is given by @xmath151 partial differential rates are defined according to @xmath152 the average values @xmath153 of the coefficient functions is given by @xmath154 where the integration over @xmath3 runs from @xmath155 to @xmath156 .
lllll @xmath157&@xmath158 ( @xmath159 ) & @xmath158 ( @xmath160 ) & @xmath161 ( @xmath159 ) & @xmath161 ( @xmath160 ) + @xmath162&@xmath163&@xmath164&@xmath165&@xmath166 + @xmath167&@xmath163&@xmath168&@xmath165&@xmath169 + @xmath170&@xmath171&@xmath172&@xmath173&@xmath174 + @xmath175&@xmath176&@xmath177&@xmath178&@xmath179 + @xmath180&@xmath181&@xmath182&@xmath183&@xmath184 + @xmath185&@xmath186&@xmath187&@xmath188&@xmath189 + @xmath190&@xmath191&@xmath192&@xmath193&@xmath194 + @xmath195&@xmath196&@xmath197&@xmath196&@xmath198 + @xmath199&@xmath200&@xmath201&@xmath200&@xmath202 + @xmath203&@xmath200&@xmath204&@xmath200&@xmath205 + in table [ normw ] we present numerical results for the normalized coefficient functions @xmath206 and their averages . in columns 2 and
3 we list the values of @xmath206 for @xmath12 with zero and nonzero lepton masses .
lepton - mass effects amount to @xmath207 for the functions @xmath208 , @xmath209 for the functions @xmath210 , and only @xmath211 for the function @xmath212 .
the normalized coefficient functions @xmath213 are zero for zero lepton masses . for @xmath160
the coefficient functions @xmath214 become quite large at @xmath215 .
again , lepton - mass effects would be even larger for smaller values of @xmath3 .
compared to the @xmath116 case , lepton mass effects are smaller by approximately a factor of @xmath216 . in columns 4 and 5 of table [ normw ] we also present average values @xmath153 of the coefficient functions again for zero and nonzero lepton masses where the average is taken with regard to @xmath3 . on average lepton mass effects
can be seen to be quite small .
let us take a closer look at the @xmath74 distribution determined by the coefficient functions @xmath218 and @xmath219 .
the normalized @xmath74 distribution is given by ( see table [ tten ] ) @xmath220 contrary to the @xmath116 case one now has a contribution linear in @xmath74 which implies a nonvanishing forward - backward asymmetry .
it is interesting to note that the source of this parity - odd term in eq .
( [ fba ] ) results from a parity - conserving interaction .
consider the @xmath221 content of the currents coupling to the @xmath222 : @xmath223 and @xmath224 .
the scalar longitudinal interference contribution leading to a nonvanishing forward - backward asymmetry can be seen to result from the parity - conserving interference of the products of currents @xmath225 , @xmath226 .
we mention that a parity - violating @xmath227 coupling as discussed in appendix a would also give rise to a nonvanishing forward - backward asymmetry proportional to @xmath117 ( see eq .
( [ wjoint3 ] ) ) . in fig .
[ wthqw ] we display the normalized @xmath74 distribution for a fixed value of @xmath12 . since @xmath228 for @xmath229 gev@xmath230 , the governing feature of the distribution is described by a downward open parabola .
the convexity parameter is proportional to @xmath231 , leading to a smaller convexity for the @xmath160 distribution as can be seen in fig .
[ wthqw ] .
there is a pronounced forward - backward asymmetry in the @xmath0 mode . according to eq .
( [ cosdistr ] ) the forward - backward asymmetry of the @xmath74 distribution is given by @xmath232 since @xmath233 and @xmath234 are positive , the forward - backward asymmetry @xmath235 is negative as also shows up in fig .
[ wthqw ] .
in fact , one calculates @xmath236 for smaller @xmath3 values the forward - backward asymmetry becomes even more pronounced . in the last column of table [ tgamhw ] we list the average value of @xmath235 which is given by @xmath237 . on average ,
the forward - backward asymmetry is much smaller than for @xmath12 .
let us rewrite the single - angle @xmath74 distribution in eq .
( [ cosdistr ] ) in terms of the transverse , longitudinal , scalar and the scalar longitudinal interference contributions .
one has @xmath239\bigg\}.\end{aligned}\ ] ] integrating the differential rate ( [ uls ] ) with respect to @xmath74 , one obtains @xmath240\right\}.\ ] ] partial decay rates are accordingly defined by @xmath241 in fig .
[ dgamhw ] we display the @xmath3 dependence of the partial decay rates .
lepton - mass effects show up mainly for lower @xmath3 values between threshold and approximately @xmath242 .
for example , at @xmath12 the helicity fractions of the @xmath238 change according to @xmath243 when going from the @xmath244 modes to the @xmath0 mode .
the picture is similar to the @xmath116 case .
however , lepton - mass effects are less pronounced in the @xmath41 case . in the upper part of table [ tgamhw ] we list the total rate and the mean values of the transverse , longitudinal and scalar partial rates for the mass - zero modes and the @xmath0 mode where the mean is taken with regard to @xmath3 . in this case , lepton - mass effects reduce the total rate by @xmath245 .
the rate reduction is largest for the longitudinal rate where the rate reduction amounts to @xmath246 . on average , the rate reduction for @xmath247 is still a considerable @xmath248 while the average of the transverse rate is practically unchanged . as in the @xmath249 case ,
the loss of longitudinal rate is mainly compensated for by the appearance of the scalar rate .
our result agrees with the result @xmath250 of ref .
@xcite within @xmath251 .
again , we conclude the section by considering the case where both @xmath253 bosons are off - shell .
using again the narrow - width approximation ( [ nwa ] ) for @xmath254 also on the @xmath57 side , the exclusive off - shell off - shell rate is obtained from the double integral @xmath255 where @xmath256 denotes the rate into the exclusive leptonic modes on the @xmath57 and @xmath58 sides .
the total inclusive mode is obtained by summing over all exclusive modes including the quark antiquark modes , i.e. by setting @xmath257 .
the rate function @xmath258 in eq .
( [ offoffrate ] ) reads @xmath259 where @xmath260 and @xmath261 .
@xmath262 and @xmath263 are the spin-1 and spin-0 projections of the charged current lepton tensors listed in appendix c. when the lepton masses are taken to be zero , the rate function ( [ wrate ] ) simplifies to @xmath264 which agrees with the results in refs .
@xcite .
llllll & @xmath265&@xmath266&@xmath267 & @xmath268&@xmath269 + 5l@xmath270 + ( @xmath271 ) & @xmath272&@xmath273&@xmath274&@xmath275 & @xmath276 + ( @xmath277 ) & @xmath278&@xmath279&@xmath280&@xmath281&@xmath282 + 5l@xmath283 + ( @xmath271 ) & @xmath284&@xmath285&@xmath286&@xmath287 & @xmath288 + ( @xmath277 ) & @xmath289&@xmath290&@xmath291&@xmath292&@xmath293 + 5l@xmath294 + ( @xmath271 ) & @xmath284&@xmath285&@xmath286&@xmath295 & @xmath296 + ( @xmath277 ) & @xmath297&@xmath298&@xmath299&@xmath300&@xmath301 + in table [ tgamhw ] we have listed our numerical results for the off - shell off - shell rates and the averages of the polarization of the gauge bosons .
the off - shell off - shell rates approximately amount to two times the off - shell on - shell rates .
the reason is again that one picks up contributions from the peaking regions both on the @xmath57 side and on the @xmath58 side .
put in a different language the off - shell off - shell rate ( [ offoffrate ] ) corresponds to the sum of the two off - shell on - shell rates @xmath302 and @xmath303 .
the polarizations listed in table [ tgamhw ] are thus an average of the respective polarizations on the @xmath57 side and the @xmath58 side .
this also explains the fact that the scalar polarization is only one - half of the on - shell off - shell value since it is contributed to only by the @xmath58 side @xmath302 channel .
this is no longer true for the @xmath304 mode where the scalar rate obtains contributions from both the @xmath57 side and the @xmath58 side .
the off - shell off - shell rates slightly exceed twice the off - shell on - shell rates which provides a measure of the quality of taking the zero - width approximation on the on - shell sides .
compared to the rates for @xmath305 listed in table [ tgamhz ] , lepton - mass effects can be seen to be more than five times smaller in the @xmath306 case .
we have discussed lepton - mass effects in the rate and the angular decay distributions in the four - body decays @xmath1 and @xmath2 where the gauge bosons @xmath13 and @xmath307 are on their mass shell .
lepton - mass effects are larger for the @xmath10 mode where we find a reduction of @xmath308 in the @xmath0 rate relative to the @xmath244 rates . in the @xmath41 case
the rate reduction of the total rate is smaller . in this mode
we find a rate reduction of @xmath245 relative to the zero mass case .
differentially , the rate reduction through lepton - mass effects is significantly larger at the lower end of the @xmath3 spectrum in both cases . for both modes
we find a significant reduction of the longitudinal rate through lepton - mass effects in the lower @xmath3 region from threshold to @xmath11 . in this region the transverse
longitudinal composition of the off - shell gauge bosons is considerably changed .
the reduction of the longitudinal rate in this region is partly compensated for by a significant scalar contribution . in the charged - current case one
finds a nonvanishing forward - backward asymmetry in the @xmath74 distribution through lepton - mass induced scalar
longitudinal interference effects .
the forward - backward asymmetry can become quite large in the low-@xmath3 region .
we have also discussed the case when both gauge - bosons go off - shell .
double smearing with the appropriate breit wigner functions increases the overall rate .
lepton - mass effects become weaker in the double smearing process .
we have employed helicity methods in our analysis which has allowed us to present our analytical results for angular decay distributions and partial rates in compact form . in particular , the inclusion of lepton - mass effects in the helicity formalism is straightforward
. experimentally it will not be so simple to identify the @xmath0 modes in the four - lepton decays of the higgs . in this context
we mention that the detection efficiency for @xmath0 leptons in their hadronic decay channels at the lhc is being continuously improved ( see refs .
@xcite ) . nevertheless , an accurate monte carlo event generator for decays involving the @xmath0 leptons should include lepton - mass effects for which we have supplied the appropriate matrix elements in this paper .
this would e.g. be relevant for modelling @xmath309 processes as background for the search for the decay @xmath310
@xcite .
this work was supported by the estonian research council under grant no .
iut2 - 27 .
j.g.k . would like to thank a. denner , b. jger , m. rauch , h. spiesberger , b. stech and j. wang for useful discussions .
s.g.acknowledges the support by the mainz institute of theoretical physics ( mitp ) .
the helicity amplitudes @xmath312 for the transition @xmath311 are defined by @xmath313 where @xmath314 and @xmath315 are the respective polarization four - vectors on the @xmath57 side ( on - shell side ) and on the @xmath58 side ( off - shell side ) .
we shall evaluate the helicity amplitudes in the higgs rest frame with the @xmath316 direction defined by the direction of the off - shell @xmath317 boson .
one therefore has to rotate the polarization four - vectors on the @xmath57 side by @xmath318 which we indicate by the `` bar '' symbol .
the respective polarization four - vectors in the higgs rest frame are given by @xmath319 where @xmath320 , @xmath321 and @xmath322 such that @xmath323 , @xmath324 .
it is convenient to avail of the covariant representations of the longitudinal and scalar polarization four - vectors .
they read @xmath325 the helicity amplitudes can then be calculated to be @xmath326 the coefficient functions @xmath327 are written in terms of bilinear forms of the helicity amplitudes for which we choose the following abbreviations @xmath328 we sometimes refer to these bilinear forms as the double spin - density matrix elements of the gauge boson pair since the bilinear forms describe the entangled polarizations components of the gauge boson pair .
the sm values of the double density matrix @xmath115 are given by @xmath329 one notes the following sm relations @xmath330 at maximal recoil where @xmath71 and/or @xmath331 tend to zero , the dominant double spin - density matrix elements are @xmath332 . at minimal recoil @xmath333 where @xmath334 , the dominant contributions are @xmath335 while @xmath233 and @xmath336 tend to zero .
some authors prefer to use cartesian components for the transition matrix elements @xcite instead of the helicity components used by us .
the relation between the two representations is given by @xmath337 we see no particular advantages to write the angular coefficient functions in terms of their cartesian components .
battacherjee _ et al .
_ considered two additional non - sm @xmath42 coupling structures @xcite .
they write down the effective coupling structure @xmath338 the helicity components are then given by @xmath339 there are no contributions of the new coupling structures to @xmath340 .
it is clear that one now has a contribution to the difference @xmath341 resulting from the parity - violating term proportional to @xmath119 .
we calculate the helicity representation of the neutral - current lepton tensor on the off - shell @xmath58 side .
the corresponding expressions for the on - shell @xmath57 side can be obtained by setting the lepton mass to zero and replacing @xmath342 .
we work in the center - of - mass system of the lepton pair with the @xmath316 direction defined by @xmath343 which we refer to as the helicity system .
the kinematics in the helicity system is given by @xmath344 where @xmath345 .
the covariant forms of the lepton tensors read @xmath346 the total neutral current lepton tensor is composed according to @xmath347 where the neutral current is defined by @xmath348 with @xmath349 in order to calculate the helicity representation of the lepton tensors in the helicity system one needs to evaluate @xmath350 all objects referring to the helicity system are denoted by a hat symbol .
the contractions are done in the helicity system using the representations ( [ qside ] ) . using the explicit forms ( [ qside ] )
one calculates @xmath351 the ratio of helicity - flip and helicity - nonflip contributions can be seen to be given by @xmath352 and @xmath353 . the components of the off - shell lepton tensor @xmath354 in the higgs decay system ( @xmath316 axis along the @xmath355 direction ) needed in the evaluation of eq .
( [ angdis2 ] ) can be obtained by rotating the components of the lepton tensor @xmath356 defined in the helicity system according to ( double indices are summed ) @xmath357 where wigner s @xmath358 functions are given by @xmath359 the rows and columns in the spin-1 part of ( [ wigner ] ) are labelled in the order @xmath360 .
one obtains @xmath361 the corresponding expressions for the on - shell side lepton tensor @xmath362 can again be obtained by rotation . note that in this case one has @xmath363 as is evident from fig .
[ planeszz ] .
the spin-1 and spin-0 projections of the neutral lepton tensor needed in the main text are given by @xmath364 where the spin-1 and spin-0 projectors read @xmath365 similar relations hold for the @xmath57 side .
the lepton tensors are given by @xmath366 the kinematics for the off - shell @xmath58 side is given by @xmath367 ( for the on - shell @xmath57 side set @xmath368 and @xmath369 ) .
the nonvanishing components of the helicity representations of the lepton tensors can then be evaluated to be @xmath370 where @xmath371 .
note that the ratio of helicity - flip and helicity - nonflip contributions are now given by e.g.@xmath372 .
as in the neutral - current case the components of the lepton tensor @xmath373 in the higgs decay system ( @xmath316 axis along the @xmath238 direction ) needed in the evaluation of eq .
( [ angdis2 ] ) can be obtained by rotating the components of the lepton tensor @xmath374 in the helicity system according to ( double indices are summed ) @xmath375 one obtains ( the rows and columns of the matrix are ordered in the sequence @xmath376 ) @xmath377 the upper / lower signs refer to the decays @xmath378 and @xmath379 .
the corresponding expressions for the on - shell - side lepton tensor @xmath380 can again be obtained by rotation . however , in this case one has @xmath363 as is evident from fig .
[ planesww ] .
further , one has to use the lower signs in the matrices ( [ lt1 ] ) .
g. aad _ et al .
_ [ atlas collaboration ] , phys .
* b716 * ( 2012 ) 1 s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
lett . * b716 * ( 2012 ) 30 s. chatrchyan _ et al . _ [ cms collaboration ] , phys . rev
. lett .
* 110 * ( 2013 ) 081803 s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
* d89 * ( 2014 ) 092007 g. aad _ et al . _ [ atlas collaboration ] , phys .
* b726 * ( 2013 ) 120 g. aad _ et al . _ [ atlas collaboration ] , arxiv:1503.03643 [ hep - ex ] s.y .
choi , d.j .
miller , m.m .
mhlleitner and p.m. zerwas , + phys .
b553 * ( 2003 ) 61 v.a .
kovalchuk , j. exp .
phys . * 107 * ( 2008 ) 774 y. gao , a.v .
gritsan , z. guo , k. melnikov , m. schulze and n.v .
tran , + phys .
* d81 * ( 2010 ) 075022 a. de rujula , j. lykken , m. pierini , c. rogan and m. spiropulu , + phys .
* d82 * ( 2010 ) 013003 s. bolognesi , y. gao , a.v .
gritsan , k. melnikov , m. schulze , n.v .
tran and a. whitbeck , phys .
* d86 * ( 2012 ) 095031 p. avery _
et al . _ ,
rev . * d87 * ( 2013 ) 055006 y. sun , x.f .
wang and d.n .
gao , int .
j. mod .
phys . * a29 * ( 2014 ) 1450086 g. buchalla , o. cata and g. dambrosio , eur .
j. * c74 * ( 2014 ) 2798 m. beneke , d. boito and y.m .
wang , jhep * 1411 * ( 2014 ) 028 j.s .
gainer , j. lykken , k.t .
matchev , s. mrenna and m. park , + phys .
* d91 * ( 2015 ) 035011 a. menon , t. modak , d. sahoo , r. sinha and h.y .
cheng , + phys .
* d89 * ( 2014 ) 095021 b. bhattacherjee , t. modak , s.k .
patra and r. sinha , arxiv:1503.08924 [ hep - ph ] t.v .
zagoskin and a.y .
korchin , arxiv:1504.07187 [ hep - ph ] j. ellis , arxiv:1504.03654 [ hep - ph ] j.g .
krner and g.a .
schuler , z. phys .
* c46 * ( 1990 ) 93 t. gutsche , m.a .
ivanov , j.g .
krner , v.e .
lyubovitskij , p. santorelli and n. habyl , + phys .
* d91 * ( 2015 ) 074001 b.a .
kniehl and o.l .
veretin , phys .
* d86 * ( 2012 ) 053007 m.e . peskin and d.v .
schroeder , `` an introduction to quantum field theory , '' reading , usa : addison - wesley ( 1995 ) 842 p l. cappiello , o. cata , g. dambrosio and d.n .
gao , + eur .
j. * c72 * ( 2012 ) 1872 [ eur .
j. * c72 * ( 2012 ) 2208 ] s.r .
gevorkyan and m.h .
misheva , eur .
j. * c74 * ( 2014 ) 2860 j.g .
krner and g.a .
schuler , z. phys .
* c38 * ( 1988 ) 511 [ z. phys .
* c41 * ( 1989 ) 690 ] m. fischer , s. groote , j.g .
krner and m.c .
mauser , phys .
* d67 * ( 2003 ) 113008 m. fael , l. mercolli and m. passera , phys .
* d88 * ( 2013 ) 093011 s. groote , j.g .
krner and l. kaldame , `` identical particle and lepton mass effects in the decay @xmath382 '' , in preparation s. groote , j.g .
krner and p. tuvike , eur .
j. * c72 * ( 2012 ) 2177 s. groote , j.g .
krner and p. tuvike , eur .
j. * c73 * ( 2013 ) 2454 g. aad _ et al . _ [ atlas and cms collaborations ] , phys .
* 114 * ( 2015 ) 191803 k.a .
olive _ et al . _
[ particle data group collaboration ] , chin .
* c38 * ( 2014 ) 090001 m. gonzalez - alonso and g. isidori , phys .
* b733 * ( 2014 ) 359 w.y .
keung and w.j .
marciano , phys .
* d30 * ( 1984 ) 248 a. djouadi , phys .
* 457 * ( 2008 ) 1 a. denner , s. heinemeyer , i. puljak , d. rebuzzi and m. spira , + eur .
j. * c71 * ( 2011 ) 1753 a. grau , g. panchieri and r.j.n .
phillips , phys .
* b251 * ( 1990 ) 293 z. zinonos , arxiv:1409.0343 [ hep - ex ] | we consider @xmath0-lepton mass effects in the cascade decays @xmath1 and @xmath2 .
since the scale of the problem is set by the off - shellness @xmath3 of the respective gauge bosons in the limits @xmath4 and not by @xmath5 , lepton - mass effects are non - negligible for the @xmath0 modes in particular close to the threshold of the off - shell decays .
lepton - mass effects show up in the rate and in the three - fold joint angular decay distribution for the decays .
nonzero lepton masses lead to leptonic helicity - flip contributions which in turn can generate novel angular dependencies in the respective three - fold angular decay distributions .
lepton - mass effects are more pronounced in the @xmath6 mode which , in part , is due to the fact that the ratio of lepton helicity - flip / nonflip contributions in the decay @xmath7 is four times larger than in the decay @xmath8 .
overall the inclusion of @xmath0 mass effects leads to a @xmath9 reduction in the leptonic @xmath10 rate .
lepton mass effects are quite pronounced for @xmath3 values from threshold up to @xmath11 .
for example , at @xmath12 the transverse longitudinal scalar helicity composition of the off - shell @xmath13boson changes from @xmath14 to @xmath15 for the @xmath0 lepton .
this has observational consequences for the angular decay distributions of the final - state leptons .
we also briefly consider the corresponding off - shell off - shell decays @xmath16 and @xmath17 .
mitp/15 - 034 + * lepton - mass effects in the decays @xmath18 and @xmath19 * + s. berge@xmath20 , s. groote@xmath21 , j.g .
krner@xmath22 and l. kaldame@xmath23 + @xmath24 institut fr theoretische physik , rwth aachen university , 52056 aachen , germany + @xmath23 loodus- ja tehnoloogiateaduskond , fsika instituut , + tartu likool , the 4 , 51010 tartu , estonia + @xmath22 prisma cluster of excellence , institut fr physik , + johannes - gutenberg - universitt , staudinger weg 7 , 55099 mainz , germany |
early type galaxies ( etgs ) are known to populate a two - dimensional manifold in the space of their observed quantities .
this fundamental plane ( fp ) is usually expressed by a correlation among the effective radius , the mean surface brightness within , and the central velocity dispersion 0 : @xmath10 due to the small intrinsic dispersion ( @xmath11 in and 0 , and @xmath12 in ) , the fp is considered as a powerful tool to measure galaxy distances and to constrain the processes driving galaxy evolution ( e.g. @xcite ) . a relevant projection of the fp is the correlation between and , also known as kormendy relation ( hereafter kr ) : @xmath13 where is the effective radius in @xmath14 , and @xmath15 @xcite .
different studies have demonstrated that in the nearby universe bright etgs define a sharp sequence in the plane , with an intrinsic dispersion of @xmath16
@xmath17 in ( @xcite ) .
moreover , @xcite have shown that @xmath18 form two distinct families in the plane of the effective parameters : that of the bright etgs , following the kr , and another ` ordinary ' family , whose properties are more disperse and heterogeneous
. recently , ( * ? ? ?
* see also references therein ) demonstrated that the claimed dichotomy between the bright and the dwarf etgs is only apparent and that there is a continuous structural relation between the two classes .
in particular , the different behavior of bright and dwarf etgs in the plane , and the change in slope of their relations , do not imply a different formation mechanism but can be interpreted with systematic changes in the profile shape with galaxy magnitude . in order to understand the processes underlying galaxy formation and evolution , it is crucial to investigate the fp and the other relations between galaxy parameters at different look - back times . in the monolithic formation scenario @xcite , etgs form at very high redshifts in a strong burst of star formation , and the distribution of galaxies in the space of observed quantities changes with redshift due to the passive fading of their stars . in the hierarchical framework ,
bright etgs form by the progressive merging of smaller units , and the fp relations are built up with redshift .
different works have shown ( 1 ) that ( dissipation - less ) merging seems to move galaxies along the fp @xcite , and ( 2 ) that in hierarchical models , the luminosity weighted ages of stellar populations in cluster galaxies appear to be the same of those predicted for pure passive evolution ( e.g. @xcite ) .
all these facts suggest that for both scenarios dynamical effects could play a minor role in the evolution of scaling correlations .
it is expected that the fp slopes change with redshift mainly for possible differences in properties of galaxy stellar populations ( such as age and metallicity ) along the early - type sequence , while the zero point varies consistently with the evolution of a passive stellar population . in the hierarchical framework , however , it is also expected that the distribution in the space of observed quantities evolves with redshift , as merging proceeds to larger and larger sizes .
clearly , in order to analyze this subject , large samples of galaxies are needed at different look - back times . in the recent years , it has become feasible to study the scaling relations at intermediate redshifts . due to the heavy demand of observing time in velocity dispersion measurements ,
many studies have investigated pure photometric laws , like the kr .
the main effort to constrain the evolution of the kr slope with redshift has been performed by @xcite , who analyzed etgs properties in four clusters at @xmath19 and in one at @xmath20 , with an average of @xmath21@xmath22 galaxies per cluster .
these authors found that the slope is comprised between 2.2 and 3.5 both for the coma and for the distant clusters , depending on the selection criteria of the samples , and suggested , therefore , that it could not have evolved since @xmath23 .
the evolution of the kr zero point for cluster galaxies has been investigated in different works in order to perform the tolman surface brightness test for the universal expansion @xcite and to derive the luminosity evolution of etgs ( e.g. @xcite ) .
all these studies concluded that the measured change is consistent with what expected on the basis of cosmological dimming plus passive aging of stellar populations formed at high redshifts .
the kr was investigated at high redshifts for galaxies in the hubble deep field by @xcite , who found that etgs in the field seem to follow the same restframe kr also at redshifts @xmath24 . as shown by @xcite , it is possible to derive structural parameters of galaxies at intermediate redshifts by using ground based imaging . in the present work ,
we use structural parameters in the v - band restframe for a total of @xmath25 spheroids in the clusters 209 at @xmath26 , 118 at @xmath27 , and @xmath28 ( hereafter 0048 ) at @xmath0 , and investigate the evolution of the distribution in the plane up to a look - back time of @xmath29 .
surface photometry for galaxies in 209 , 118 and 0048 is published in @xcite .
one of the main challenges of previous works on the subject has been represented by selection effects , which can critically affect the determination of the kr . to work out this problem
, we introduce a new fitting procedure , which provides un - biased estimates of the kr coefficients , and allows for the first time to investigate the evolution over redshift of the whole distribution in the plane of effective parameters .
the layout of the paper is the following . in section [ sdata ] , we present the samples used for the analysis .
section [ skrel ] describes the new fitting procedure and deals with the evolution of the kr slope and the comparison of the distributions at different redshifts .
section [ slev ] shows the luminosity evolution inferred by the kr .
conclusions are drawn in section [ sconc ] . in the following
, we will assume the cosmology @xmath30 , @xmath31 and @xmath32 .
with these parameters , the age of the universe is @xmath33 , and the redshifts @xmath34 , and @xmath0 correspond to look - back times of @xmath35 , and @xmath36 respectively .
photometry for the three clusters of galaxies 209 ( @xmath26 ) , 118 ( @xmath27 ) , and 0048 ( @xmath0 ) resides on data collected at the eso new technology telescope ( ntt ) and at the eso very large telescope ( vlt ) .
the data relevant for the present study include images in the r band for 209 and 118 , and in the i band for 0048 , all corresponding approximately to v - band restframe . for 118 and 0048 , galaxies were selected according to the photometric redshift technique ( see @xcite , and references therein ) , while for 209 we considered the sample of spectroscopically confirmed cluster members from @xcite complemented with photometric data ( see below ) .
the relevant information on the samples are summarized in table [ data ] .
structural parameters were derived as described in @xcite , by fitting galaxy images with sersic models convolved with suitable representations of the psf .
uncertainties on structural parameters were estimated as described in @xcite , and typically amount to @xmath37 and @xmath38 for 209 , @xmath39 and @xmath40 for 118 , and to @xmath41 and @xmath42 for 0048 . in order to study the kr , we selected the population of spheroids on the basis of the shape of the light profile , as parametrized by the sersic index .
we classified as spheroids the galaxies with @xmath43 , corresponding to objects with a bulge fraction greater than @xmath44 ( see @xcite ) .
for each cluster we considered all the galaxies brighter than a given magnitude limit @xmath45 .
few objects with small radii and poorly determined structural parameters were excluded by selecting only galaxies with greater than a given surface brightness limit @xmath46 .
the completeness of the spectroscopic sample of 209 is about @xmath47 at @xmath48 , and falls to zero at @xmath49 .
a magnitude complete sample was obtained by adding to the spectroscopic data all the spheroids within the colour magnitude sequence of 209 ( see @xcite ) down to @xmath50 .
we point out that the present samples constitute the largest data set ( @xmath51 ) of cluster galaxies , for which the kormendy relation is investigated up to a look - back time of @xmath52 . in order to compare the distributions in the plane at intermediate redshifts with that of nearby galaxies
, we took advantage of the large sample of etgs in the coma cluster by ( * ? ? ?
* hereafter jfk95a )
. structural parameters for @xmath53 galaxies were taken in jfk95a , who fitted the growth curve of aperture magnitudes with de vaucouleurs @xmath54 profiles and corrected the estimated parameters for seeing effects .
the typical uncertainties on and amount to @xmath55 and @xmath56 , respectively ( see table 11 of jfk95a ) .
since it was not possible to select galaxies in the jfk95a sample by using sersic indices as for the other clusters , we adopted the relation between sersic index and central velocity dispersion recently found for nearby galaxies by @xcite : @xmath57 .
we considered the sample of coma etgs with available velocity dispersion from ( * ? ? ?
* hereafter jfk95b , @xmath58 ) and from ( * ? ? ?
* hereafter j99 , @xmath59 ) . among these galaxies ,
@xmath60 objects have surface photometry from jfk95a . according to @xcite ( figure 1 ) , we selected galaxies with @xmath61 , which corresponds to @xmath62 , resulting in a sample of @xmath63 objects .
we also excluded from the calculation of the fitting coefficients the three galaxies with largest size in the jfk95a sample : due to the presence of extended halos the use of a de vaucouleurs model can give very different parameters with respect to the estimates obtained by the sersic models .
the final sample consists in @xmath64 galaxies brighter than @xmath65 , where @xmath66 is the thuan
gunn r - band magnitude band and the v - band restframe at the redshift is not relevant for the present study ] .
we note that , due to the previous selections , the coma sample does not constitute a magnitude complete sample of galaxies .
this point will be further addressed in section [ fitkr ] . for 209 , 118 , and 0048 , mean surface brightnesses were corrected for galactic extinction following @xcite , while the cosmological dimming was not removed . for the sample ,
since both corrections were applied by jfk95a , we added the term @xmath67 to the values .
the diagrams for , 209 , 118 , and 0048 are shown in figure [ krel ] . for each cluster , spheroids follow a sharp , well defined kr .
since selection effects can strongly affect the estimate of the kr coefficients ( see @xcite ) , we fitted the sequences by introducing a modified least square procedure ( hereafter @xmath68 , see appendix [ appa ] ) , which corrects the bias due to the different completeness cuts in magnitude and of each sample .
the fitting coefficients were derived by minimizing the rms of the residuals with respect to and ( and fits ) , and by applying the bisector regression ( , see @xcite ) , which is more stable and effective with respect to the other fitting methods . in order to illustrate the reliability of the adopted procedure
, we applied different magnitude cuts , @xmath69 , to the sample of 118 , which has the largest number of galaxies , and for each value of @xmath69 we derived the slope and the zero point of the kr by using the @xmath68 and the ordinary least square ( @xmath70 ) fits . in figure [ bias_fitkr ]
we show the variations of @xmath4 and @xmath71 as a function of @xmath69 for the bisector fit .
similar results are obtained for the and the regressions .
the completeness cut affects significantly the ols method , producing the well known malmquist bias : for lower values of @xmath69 the slope of the kr is overestimated while the value of @xmath71 decreases systematically . on the contrary ,
it is evident that the @xmath68 procedure removes efficiently the systematic trends of the @xmath70 method , providing un - biased estimates of the kr coefficients .
this implies that the @xmath68 fit allows a straightforward comparison of the slope and of the zeropoint of the kr , ruling out any effect related to the different selection criteria of the samples .
since the selection criteria for the coma cluster are different with respect to those of the other clusters , we also verified that the fitting results for this sample are not affected by the cuts in magnitude and velocity dispersion . according to the scatter of the relation between velocity dispersion and sersic index by @xcite dispersion of @xmath72 , we varied the velocity dispersion cut within @xmath73 and re - derived the @xmath68 coefficients .
it turned out that the slope @xmath4 of the kr changes by few percents , while the corresponding variations of the zeropoint are within @xmath74 .
these values are fully consistent with the estimates of @xmath4 and @xmath71 that are given in sections [ sbev ] and [ slev ] .
a similar result was obtained by changing the magnitude limit of the sample down to @xmath75 .
it is worth to be noted that the values we obtained for the coefficients of the kr at @xmath76 are in very good agreement ( see sections [ sbev ] and [ slev ] ) with those of previous works for bright nearby etgs , spanning a much wider luminosity range ( see e.g. @xcite ) .
the selection effects on the local reference sample do not affect , therefore , the present analysis .
another important point is the correlation between the uncertainties on galaxy parameters : for the effective parameters , we have @xmath77 ( see jfk95a ) , where the value of @xmath78 depends on the methods used to derive and , and the corresponding uncertainties . in order to estimate the range of possible values of @xmath78 , we compared the uncertainties on and for samples of galaxies from different sources ( @xcite ) : we found @xmath79 . by using numerical simulations ( see @xcite ) and by comparing repeated measurements of and ( see @xcite ) , we obtained @xmath80 for the parameters of 209 , 118 , and 0048 , in agreement with the quoted range of @xmath78 values .
the correlation between and and the typical values of the uncertainties on the effective parameters are shown in figure [ krel ] . the effect on the kr coefficients
was estimated by using numerical simulations of the diagrams , by a procedure analogous to that described in @xcite .
we found that the bias on the kr coefficients is negligible , amounting at most to @xmath81 for 118 .
this is in agreement with what found by @xcite for the relation of nearby bright ellipticals and spiral bulges .
the slopes of the krs are reported in table [ krslope ] , while the variation of @xmath4 with redshift is shown in figure [ bev ] .
the uncertainties were estimated by the bootstrap method and amount to @xmath82@xmath83 for 209 , 118 , and 0048 .
we note that the uncertainty for the sample is significantly higher ( @xmath84 @xmath85 ) due to the smaller range of .
figure [ bev ] clearly shows that the values of @xmath4 are fully consistent for each pair of clusters , except for the and slopes of the sample , which are consistent with the other values at a lower significance level ( @xmath86 ) .
this is due to the fact that the and methods are more sensitive , with respect to the method , to the range of used in the fit , that is fully consistent with the value in table [ bev ] . ] .
we conclude , therefore , that no significant evolution in the slope of the kr is found up to @xmath87 .
we also note that the weighted means of @xmath4 reported in table [ krmslope ] for the various fitting procedures are in very good agreement with values found by previous studies at @xmath88 : e.g. @xmath89 ( @xcite ) , @xmath90 ( @xcite ) and @xmath91 ( @xcite ) .
the evolution of the kr slope with redshift , @xmath92 , carries information on the variation in the properties of the mean stellar populations ( sps ) as a function of galaxy radius : @xmath93 where @xmath94 is the luminosity evolution between @xmath0 and @xmath95 at a given . in appendix
[ appb ] , we use the gissel98 synthesis code ( @xcite ) to derive an approximated analytic expression for @xmath96 as a function of the age @xmath97 and the metallicity @xmath98 of the galaxy sps .
we consider luminosity evolution in the v band , since ( 1 ) the filters relative to each sample approximate the v - band restframe and ( 2 ) internal colour gradients of spheroids are known to be very small in the optical wavebands and not to evolve significantly at intermediate redshifts ( see @xcite ) . for models with formation redshift @xmath99}$ ] and metallicity @xmath100z_\odot}$ ] , which are suitable to describe the population of spheroids ( see also section [ slev ] ) , we show that @xmath96 depends mainly on @xmath97 , i.e. @xmath101 , so that eq . [ delb ] can be approximated as : @xmath102 where @xmath103 is computed as described in appendix [ appb ] . assuming that the maximum allowed variation of @xmath4 is @xmath104 ( see table [ krmslope ] ) , we obtain the relation @xmath105 for a ssp , @xmath106~mag / gyr}$ ] , and the value of @xmath107 turns out to be in the range @xmath108 $ ] .
this implies that the absolute relative variation of @xmath97 per decade of galaxy radius must be smaller than @xmath109 ( @xmath110 ) .
as shown in appendix [ appb ] , a lower limit is obtained for sp models with a protracted sfr . for an exponential sfr with time scale @xmath111 , we find @xmath112~mag / gyr}$ ] and the relative variation of @xmath97 is constrained to be smaller than @xmath113 . since the values of @xmath4 are the same at all redshifts , we adopted as common slope in the kr fit the value relative to the regression , @xmath114 , and computed for each sample the zeropoint @xmath71 and the scatter @xmath115 , defined by the rms of the residuals @xmath116 in : @xmath117 .
the uncertainties on @xmath71 and @xmath118 were obtained by numerical simulations , taking also into account the measurement error on the adopted value of the slope .
we estimated the intrinsic dispersion of the relations ( ) by subtracting in quadrature to @xmath119 the amount of dispersion on @xmath116 due to the measurement errors on and . to this aim
, we took into account the typical uncertainties on and ( see section [ sdata ] ) , and their correlation ) .
for the cluster , we adopted the covariance term used for 0048 , whose structural parameters have similar uncertainties . ] .
the values are given in table [ krpar ] .
it is clear from the table that the kr has significant intrinsic dispersion also at intermediate redshifts ( @xmath120 in ) , and that this dispersion does not change with redshift .
the intrinsic dispersion of the kr is consistent with that reported by @xcite and @xcite for nearby galaxies , and is mainly due to neglecting velocity dispersions in the fp relation ( eq .
[ fpeq ] ) . to compare the distributions in the plane , we first shifted the values of for 209 , 118 , and 0048 by matching the zeropoints of the krs , and then applied to each sample the same cuts in magnitude and .
since the @xmath68 fit provides un - biased estimates of the kr coefficients , this procedure is not affected by the different selection criteria of each sample .
the cluster was not considered in order to analyze only the samples with structural parameters derived by the same procedure ( see below ) . the combined diagram for 209 , 118 , and 0048
is shown in figure [ krtot ] . for each cluster , we derived the mean value , @xmath121 , and the standard deviation , @xmath122 , of the distribution by applying the bi - weight statistics ( e.g. @xcite ) , which has the advantage to minimize the effect of outliers .
as shown in table [ lrecof ] , the values of @xmath123 and @xmath122 are fully consistent ( within @xmath124 ) for each pair of clusters among the clusters .
in fact , by subtracting in quadrature the error on @xmath125 from @xmath122 , we obtain the following estimates of the standard deviations : @xmath126 , @xmath127 and @xmath128 , for 209 , 118 , and 0048 , respectively . ] . in order to test the presence of differences in the distributions of effective radii
, we also applied the kolmogorov
smirnov ( ks ) test .
the probability that the observed samples come from the same parent distribution turned out to be @xmath129 for 209 and 118 , @xmath130 for 209 and 0048 , and @xmath131 for 118 and 0048 , indicating that the distribution of galaxy sizes does not change significantly over the explored redshift interval . in order to verify the presence of possible differences in the diagram among the clusters
, we used the two - dimensional ks test ( @xcite ) .
the test was repeated by using different shifts for each cluster to take into account the uncertainties on @xmath71 reported in table [ krpar ] . comparing 118 with the other two clusters , we found probabilities of the ks statistics between @xmath132 and @xmath133 , while a larger probability ( @xmath134 ) was obtained by comparing 209 and 0048 .
we point out that the last result is of particular interest , since the measurement uncertainties on the parameters of 209 and 0048 are similar and are comparable with the intrinsic dispersion of the kr .
since the error bars for the 118 data are significantly larger , we also performed the ks test by adding further scatter to the diagram of 209 and 0048 in order to mimic the measurement uncertainties on the effective parameters at @xmath135 . in this case
, we obtained higher probabilities , between @xmath136 and @xmath137 , that the sample of 118 is drawn from the same parent distribution of the other two clusters . by considering also the cluster in the comparison of the distributions , we obtain @xmath138 and @xmath139 , while the ks tests for both the one - dimensional and two - dimensional case give probabilities between @xmath140 and @xmath132
the fact that the ks probabilities and the value of @xmath141 are lower with respect to those obtained for the other distributions can be explained by the different procedure by which the structural parameters of the cluster are derived , and is therefore not relevant for the present analysis .
the previous results , together with those of sections [ sbev ] and [ ssev ] , indicate that the distribution in the plane for cluster galaxies does not vary significantly at least back to @xmath142 .
the evolution of the zeropoint of the kr over redshift is determined ( 1 ) by the variation of the distribution of galaxy sizes and of the kr slope with look - back time , ( 2 ) by the effect of the cosmological dimming on , ( 3 ) by the luminosity evolution of galaxy stellar populations , ( 4 ) by the values of cosmological parameters .
point ( 1 ) is ruled out by the results of sections [ sbev ] and [ sdev ] . since the investigation of point ( 4 ) would require a larger number of cluster samples
, we will focus the discussion on points ( 2 ) and ( 3 ) .
we start to note that , as for the kr slope , the value of @xmath71 obtained by the @xmath143 fit for the coma sample is consistent with that found by previous studies at @xmath76 .
for example , if we adopt the kr slope obtained by ( * ? ? ?
* hereafter hos87 ) in the @xmath144 band for a sample of 97 bright nearby early types ( @xmath145 ) ,
we obtain @xmath146 for the sample , which is in very good agreement with that of hos87 reported to our cosmology : @xmath147 .
suitable corrections were applied to the data in order to convert the values to the same restframe waveband ( v ) for each cluster .
we used the gissel98 synthesis code ( @xcite ) to construct galaxy templates with different age @xmath97 ( referred to @xmath95 ) , metallicity @xmath98 and time scale @xmath148 of star formation . in order to consider a wide range of possible spectral types describing the galaxies in our sample , we chose the following range of parameters : @xmath149~gyr}$ ] , @xmath150z_\odot}$ ] and @xmath151~gyr}$ ] .
the mean values of the derived corrections and the relative uncertainties are shown in table [ kcor ] .
we notice that the uncertainties are very small ( @xmath152 ) , i.e. the dependence on the spectral type is negligible , due to the fact that the observed wavebands approximate closely the v - band restframe for each cluster
. we also verified that using ssp spectra by @xcite gives values fully consistent with those of table [ kcor ] .
this proves that it is correct to apply the same corrections to all the galaxies in each cluster .
the v - band zeropoints ( @xmath153 ) of the kr were obtained by adding the values reported in table [ kcor ] to the values of @xmath71 ( table [ krpar ] ) , and the corresponding error budgets were estimated by adding in quadrature the uncertainties on the corrections with @xmath154 . in figure [ lev ]
we show the relation between @xmath153 and @xmath155 .
the observed points are well described by a linear relation @xmath156 .
a least square fit gives @xmath157 , or equivalently : @xmath158 , where @xmath159 is the mean surface brightness within in linear units .
as shown by @xcite , this excludes the relation @xmath160 , which would hold in the case of a non - expanding universe .
we compare the relation between @xmath153 and @xmath155 with a no evolution sequence , that is with a pure tolman signal @xmath161 , and with models that include both cosmological dimming and passive luminosity evolution ( see figure [ lev ] ) .
these models were constructed by the gissel98 synthesis code as follows .
we considered galaxy templates with solar metallicity @xmath162 , with different formation redshifts ( @xmath163 ) , and with two time scales of star formation : @xmath164 , which gives a suitable description of the evolution of etg integrated colours , and @xmath165 , which reproduces the properties of a simple stellar population .
changing the value of @xmath98 in the range @xmath166@xmath167 does not change significantly the results .
the models were arbitrarily scaled to fit the data . according to what found by previous studies ( see @xcite ) ,
the trend of @xmath168 with redshift is well described by the superposition of the tolman signal with pure ( passive ) luminosity evolution .
the formation redshift of the stellar populations is constrained to be higher than @xmath169 with a confidence level @xmath170 , in agreement with what found by previous studies of the kr ( @xcite and @xcite ) and of the fp ( @xcite ; @xcite ; @xcite ; @xcite ) .
we have studied the evolution of the optical , v - band restframe , kormendy relation in the range @xmath171 , by using ground - based effective parameters derived in previous works for the clusters of galaxies 209 ( @xmath26 ) , 118 ( @xmath27 ) , and 0048 ( @xmath0 ) , and published data for the coma cluster ( @xmath172 ) .
galaxies were selected by the photometric redshift technique for 118 and 0048 and by spectroscopic redshifts complemented with photometric data for 209 .
we introduce a new fitting procedure , the @xmath68 fit , which provides un - biased estimates of the kr coefficients .
this allows , for the first time , to investigate the evolution of the slope and of the scatter of the kr , as well as of the whole distribution in the plane of effective parameters .
the results of the present work can be summarized as follows .
@xmath173 : : spheroids define a tight sequence in the plane of the effective parameters at least back to @xmath174 .
@xmath173 : : the slope @xmath4 and the intrinsic dispersion turn out to be consistent at all redshifts .
the mean values of @xmath4 and are @xmath175 and @xmath176 , respectively .
@xmath173 : : the fact that the kr slopes are the same among all the clusters constrains the possible change in the formation epoch of galaxy stellar populations per decade of radius to be smaller than @xmath109 for ssp models , and smaller than @xmath113 for populations having an exponential sfr with @xmath164 . @xmath173 : : the mean value and the width of the distribution of galaxy sizes are consistent for each pair of clusters .
moreover , the kolmogorov smirnov tests indicate that the whole distribution in the plane do not vary over redshift .
@xmath173 : : in agreement with previous studies , the evolution of the kr zeropoint is explained by the superposition of the tolman signal and the passive evolution of stellar populations with high formation redshift , @xmath9 .
these results imply that , whatever the mechanism driving galaxy evolution is , ( 1 ) it has already built up the kormendy relation at @xmath0 and ( 2 ) it does not affect significantly the properties of spheroids in the plane of effective parameters from that redshift down to @xmath177 .
the fact that the kr slope does not evolve with redshift sets constrains on the differential luminosity evolution along the early - type sequence , and , therefore , on the properties of galaxy stellar populations as a function of galaxy radius .
these results should be taken into account by any model aimed at reproducing the physical processes underlying formation and evolution of spheroids .
we are grateful to r. de carvalho and a. iovino for the helpful discussions .
we thank a. mercurio for providing us with the photometry of the cluster of galaxies 209 .
michele massarotti is partly supported by a ` miur - cofin ' grant .
we consider the statistical model : @xmath178 where @xmath179 and @xmath180 are two random variables describing the distributions of and ( or viceversa ) and a and b are the zero point and the slope of the kr . in order to derive a and b
, we have to take into account the constrains on @xmath179 and @xmath180 arising from the selection cuts . by considering the case of a single cut
, we have the following constrain : @xmath181 where x and y are the outputs of @xmath179 and @xmath180 , while the coefficients @xmath182 and @xmath183 describe the selection cut .
the direction of the inequality in eq .
[ a2 ] is chosen arbitrarily . in the case of the magnitude limit @xmath45 , with @xmath184 and @xmath185 , we have @xmath186 and @xmath187 .
this follows from the relation @xmath188 @xmath189 , where @xmath190 is the total magnitude , and from the constrain @xmath191 . for a fixed value of @xmath192
, the probability to find a point in the interval @xmath193 is given by @xmath194 \cdot f(c_1+c_2 \cdot x - y ) d \ !
y } \label{pmod}\ ] ] where f is a step function , and the coefficient k is derived by imposing @xmath195 .
. [ pmod ] can be directly generalized to the case of multiple cuts and to a more complex shape of the selection function f. we write the likelihood @xmath196 , and derive the coefficients in eq .
[ lmod ] by minimizing the function @xmath197 with respect to a and b. we notice that in the ordinary least square fit the function f is absent in eq .
therefore , the term k does not depend on @xmath198 and @xmath199 and the minimization of @xmath200 reduces to solving a coupled system of linear equations in a and b.
spectra of galaxy templates were constructed by the gissel98 synthesis code , considering models with a scalo imf and an exponential sfr , @xmath201 .
we considered ssp spectra and models having a protracted sfr , @xmath202 , with metallicity @xmath203 and age @xmath204~gyr}$ ] .
the range of @xmath97 was chosen in order to describe the properties of galaxy templates having formation redshift @xmath99}$ ] ( corresponding to a look - back time @xmath205~gyr}$ ] ) , from @xmath0 to @xmath95 . for both models
, we found that the dependence of the v - band magnitude at @xmath206 on @xmath97 and @xmath98 is well described by a second - order polynomial fit : @xmath207^i + \sum_{i=1,2 } d_i \cdot \left [ \log z \right]^i + cost , } \label{b1}\ ] ] with a rms of @xmath208 .
the restframe v - band magnitude at @xmath0 is obtained by substituting @xmath97 with @xmath209 in eq .
[ b1 ] , where @xmath210 is the look - back time corresponding to @xmath211 .
we point out that the presence of cross terms in eq .
[ b1 ] , such as @xmath212 , reduces the rms of the fit only marginally , by less than @xmath213 , and that , therefore , the luminosity evolution @xmath214 depends mainly on @xmath97 , that is @xmath101 .
the derivative of @xmath215 with respect to @xmath97 can be computed analytically from eq .
[ b1 ] and is given by the following equation : @xmath216 + 2 c_2 \cdot \left [ \frac{t_f \cdot \log ( t_f - t_{0.64})}{t_f - t_{0.64 } } - \log t_f \right ] } .\ ] ] this relation was used to compute the range of values of @xmath217 in section [ sbev ] .
akritas , m.g . , and bershady , m.a .
1996 , , 470 , 706 barger , a.j .
, aragn - salamanca , a. , smail , i. , ellis , r.s . , couch , w.j . ,
dressler , a. , oemler , a. , poggianti , b.m . , and sharples , r.m .
1998 , , 501 , 522 beers , t.c . ,
flynn , k. , and gebhardt , k. 1990 , , 100 , 32 bruzual , g.a . , and charlot , s. 1993 , , 405 , 538 busarello , g. , merluzzi , p. , la barbera , f. , massarotti , m. , and capaccioli , m. 2002 , , 389 , 787 buzzoni , a. 1989 , , 71 , 817 capaccioli , m. , caon , n. , and donofrio , m. 1992 , , 259 , 323 capelato , h.v .
, de carvalho , r.r . , and carlberg , r.g .
1995 , , 451 , 525 dantas , c.c . ,
capelato , h.v . ,
ribeiro , a.l.b .
, and de carvalho , r.r .
2003 , , 340 , 398 fasano , g. , cristiani , s. , arnouts , s. , and filippi , m. 1998 , , 115 , 1400 fasano , g. , and franceschini , a. 1987 , , 225 , 155 graham , a. 2002 , , 334,859 graham , a. , and guzmn , r. 2003 , , in press ( astro - ph/0303391 ) hamabe , m. , and kormendy , j. 1987 , in : _ structure and dynamics of elliptical galaxies , iau symp .
, no.127 _ , p.379 , ed .
de zeeuw , t. , reidel , dordrecht .
hoessel , j.g . ,
oegerle , w.r . , and schneider , d.p .
1987 , , 94 , 1111 jrgensen , i. , franx , m. , and kjrgaard , p. 1995a
, , 273 , 1097 jrgensen , i. , franx , m. , and kjrgaard , p. 1995b
, , 276 , 1341 jrgensen , i. , franx , m. , and kjrgaard , p. 1996
, , 280 , 167 jrgensen , i. 1999 , , 306 , 833 , 607 jrgensen , i. , franx , m. , hjorth , j. , and van dokkum , p. g. 1999 , , 308 , 833 kauffmann , g. 1996 , , 281 , 487 kelson , d.d . , illingworth , g.d . , van dokkum , p.g . , and franx , m. 2000 , , 531 , 184 kelson , d.d . , van dokkum , p.g . , franx , m. , illingworth , g.d . , and fabricant , d. 1997 , , 478 , 13 kjrgaard , p. , jrgensen , i. , and moles , m. 1993 , , 418 , 617 kormendy , j. 1977 , , 218 , 333 la barbera , f. , busarello , g. , and capaccioli , m. 2000 , , 362 , 851 la barbera , f. , busarello , g. , merluzzi , p. , massarotti , m. , and capaccioli , m. 2002 , apj , 571 , 790 la barbera , f. , merluzzi , p. , iovino , a. , busarello , g. , massarotti , m. , and capaccioli , m. 2003 , , 399 , 899 la barbera , f. , busarello , g. , massarotti , m. , merluzzi , p. , and mercurio , a. 2003 , , in press ( astro - ph/0211478 ) larson , r.b .
1974 , , 166 , 585 lubin , l.m . , and sandage , a. 2001 , , 121 , 2289 lubin , l.m . , and sandage , a. 2001 , , 122 , 1071 lubin , l.m . , and sandage , a. 2001 , , 122 , 1084 lucey , j.r . ,
bower , r.g . , and ellis , r.s .
1991 , , 249 , 755 massarotti , m. , iovino , a. , buzzoni , a. , and valls - gabaud , d. 2001 , , 380 , 425 mercurio , a. , girardi , m. , boschin , w. , merluzzi , p. , and busarello , g. 2002 , , 397 , 431 mercurio , a , massarotti , m. , merluzzi , p. , girardi , m. , busarello , g. , and la barbera , f. 2003 , , submitted ( astro - ph/0303598 ) mobasher , b. , guzman , r. , aragon - salamanca , a. , and zepf , s. 1999 , , 304 , 225 pahre , m.a . , djorgovski , s.g . , and de carvalho , r.r .
1996 , 456 , 79 pahre , m.a . ,
djorgovski , s.g . ,
de carvalho , r.r .
2001 , , 276 , 983 saglia , r. p. , bender , r. , and dressler , a. 1993 , , 279 , 75 saglia , r.p .
, bertschinger , e. , baggley , g. , burstein , d. , colless , m. , roger , l.d . ,
mcmahan , r.k . ,
jr . , and wegner , g. 1997 , , 109 , 79 sandage , a . , and lubin , l.m .
2001 , , 121 , 2271 sandage , a. , and perelmuter , j .- m .
1991 , , 370 , 455 schlegel , d.j .
, finkbeiner , d.p . , and davis , m. 1998 , , 500 , 525 scodeggio , m. , giovanelli , r. , and haynes , m. p. 1997
, , 113 , 101 scodeggio , m. , gavazzi , g. , belsole , e. , pierini , d. , and boselli , a. 1998 , , 301 , 1001 van dokkum , p.g . , franx , m. , kelson , d.d . , illingworth , g.d . , fisher , d. , and fabricant , d. 1998 , , 500 , 714 van dokkum , p.g . ,
franx , m. , kelson , d.d . , and illingworth , g.d .
1998 , 504 , 17 ziegler , b.l . ,
saglia , r.p . , bender , r. , belloni , p. , greggio , l. , and seitz , s. 1999 , , 346 , 13 lcccccc coma & 0.023 & @xmath218 & 90 & ... & 15.3 & jfk95a , jfk95b , j99 + 209 & 0.21 & @xmath219 & 81 & 18.6 & 20.1 & mgb03 , mmm03 , lbm03 + 118 & 0.31 & @xmath219 & 101 & 19.1 & 20.8 & bml02 , lbm02 + 0048 & 0.64 & @xmath220 & 46 & 18.9 & 22.0 & lmi03 , lbm03 + | we study the evolution of the relation between half - light ( effective ) radius , , and mean surface brightness , , known as kormendy relation , out to redshift @xmath0 in v - band restframe on the basis of a large sample of spheroidal galaxies ( @xmath1 ) belonging to three clusters of galaxies .
the present sample constitutes the largest data set for which the kormendy relation is investigated up to a look - back time of @xmath2 ( @xmath3 ) . a new fitting procedure , which suitably accounts for selection criteria effects , allows for the first time to study the trend of the slope ( @xmath4 ) and of the intrinsic dispersion ( ) of the kormendy relation , and the properties of the whole distribution in the plane as a function of look - back time . the slope @xmath4 of the relation does not change from @xmath0 to the present epoch : @xmath5 , implying a tight constraint of @xmath6@xmath7 on the variation of the stellar formation epoch along the sequence of spheroidal galaxies per decade of radius .
the intrinsic dispersion of the relation , @xmath8 , does not vary with redshift and the distribution of galaxy sizes as well as the distribution in the plane of the effective parameters do not vary among the clusters , as proven by the kolmogorov smirnov tests .
we conclude that , whatever the mechanism driving galaxy evolution is , it does not affect significantly the properties of bright galaxies in the plane at least since @xmath0 .
the evolution of the zeropoint of the kormendy relation is fully explained by the cosmological dimming in an expanding universe plus the passive luminosity evolution of stellar populations with high formation redshift ( @xmath9 ) . |
let @xmath4 denote the space of ( ordered ) partitions of @xmath5 , that is @xmath6 by _ size - biased _ sampling according to a point @xmath7 we mean picking the @xmath8-th part @xmath9 with probability @xmath9 .
the starting point for our study is the following markov chain on @xmath4 , which we call a _ coagulation - fragmentation _ process : size - bias sample ( with replacement ) two parts from @xmath10 .
if the same part was picked twice , split it ( uniformly ) , and reorder the partition . if different parts were picked , merge them , and reorder the partition .
we call this markov chain the
_ basic chain_. we first bumped into it in the context of triangulation of random riemann surfaces @xcite .
it turns out that it was already considered in @xcite , in connection with `` virtual permutations '' and the poisson - dirichlet process .
recall that the poisson - dirichlet measure ( with parameter 1 ) can be described as the probability distribution of @xmath11 on @xmath4 obtained by setting @xmath12 , @xmath13 , and reordering the sequence @xmath14 , where @xmath15 is a sequence of i.i.d.uniform[0,1 ] random variables .
tsilevich showed in @xcite that the poisson - dirichlet distribution is an invariant probability measure for the markov chain described above , and raised the question whether such an invariant probability measure is unique . while we do not completely resolve this question , a corollary of our results ( c.f .
theorem [ theo - uniq ] ) is that the poisson - dirichlet law is the unique invariant measure for the basic chain which satisfies certain regularity conditions . of course
, the question of invariant probability measure is only one among many concerning the large time behavior of the basic chain . also , it turns out that one may extend the definition of the basic chain to obtain a poisson - dirichlet measure with any parameter as an invariant probability measure , generalizing the result of @xcite .
we thus consider a slightly more general model , as follows . for any nonnegative sequence @xmath16 ,
let @xmath17 , the @xmath18 norm of @xmath19 , and @xmath20 .
set @xmath21 and @xmath22 and @xmath23 . unless otherwise stated
, we equip all these spaces with the topology induced from the product topology on @xmath24 . in particular
, @xmath25 is then a compact space . for a topological space @xmath26 with borel @xmath3-field @xmath27
we denote by @xmath28 the set of all probability measures on @xmath29 and equip it with the topology of weak convergence .
@xmath30 denotes the space of all ( nonnegative ) measures on @xmath29 .
define the following two operators , called the _
merge _ and _ split _ operators , on @xmath31 , as follows : @xmath32 note that the operators @xmath33 and @xmath34 preserve the @xmath18 norm .
let @xmath35)$ ] be a probability measure on @xmath36 $ ] ( the _ splitting measure _ ) . for @xmath37 and @xmath38 $ ] , we then consider the markov process generated in @xmath25 by the kernel @xmath39 it is straightforward to check ( see lemma [ cont ] below ) that @xmath40 is feller continuous .
the basic chain corresponds to @xmath41 $ ] , with @xmath42 .
it is also not hard to check ( see theorem [ exists ] below ) that there always exists a @xmath43-invariant probability measure @xmath44 .
basic properties of any such invariant probability measure are collected in lemma [ schteim ] and proposition [ prop ] .
our first result is the following characterization of those kernels that yield invariant probability measures which are supported on finite ( respectively infinite ) partitions . to this end , let @xmath45 be the set of finite partitions . *
( support properties ) * [ theo - duo ] for any @xmath43-invariant @xmath44 , @xmath46=1&\mbox{\ if\ } & \int\frac{1}{x}\ d{\sigma}(x)\ < \
\infty\qquad\mbox{and}\\ \mu[s]=0&\mbox{\ if\ } & \int\frac{1}{x}\ d{\sigma}(x)\ = \ \infty.\end{aligned}\ ] ] transience and recurrence criteria ( which , unfortunately , do not settle the case @xmath41 $ ] ! ) are provided in the : * ( recurrence and transience ) * [ rectrans ] the state @xmath47 is positive recurrent for @xmath43 if and only if @xmath48 .
if however @xmath49}\ dx<\infty\ ] ] then @xmath50 is a transient state for @xmath43 .
we now turn to the case @xmath41 $ ] . in order to define invariant probability measures in this case ,
set @xmath51 for each @xmath52 consider the poisson process on @xmath53 with intensity measure @xmath54 which can be seen either as a poisson random measure @xmath55 on the positive real line or as a random variable @xmath56 taking values in @xmath57 whose distribution shall be denoted by @xmath58 , with expectation operator @xmath59 .
( indeed , @xmath60while @xmath61)=0 $ ] , and thus @xmath62 a.s . ) .
a useful feature of such a poisson process is that for any borel subset @xmath63 of @xmath53 with @xmath64 , and conditioned on @xmath65 , the @xmath66 points in @xmath63 are distributed as @xmath66 independent variables chosen each according to the law @xmath67 .
the poisson - dirichlet measure @xmath68 on @xmath4 is defined to be the distribution of @xmath69 . in other words , @xmath70 . in the case
@xmath71 it coincides with the previously described poisson - dirichlet measure .
see @xcite , @xcite and @xcite for more details and additional properties of poisson - dirichlet processes .
we show in theorem [ theo - uniq ] below that , when @xmath41 $ ] , for each choice of @xmath72 there is a poisson dirichlet measure which is invariant for @xmath40 .
we also show that it is , in this case , the unique invariant probability measure in a class @xmath73 , which we proceed to define .
set @xmath74 and denote by @xmath75 the set of real valued functions on @xmath76 that coincide ( [email protected] . ) with a function which has a real analytic extension to some open neighborhood of @xmath78 .
( here and throughout , @xmath79 denotes the @xmath80-dimensional lebesgue measure ; all we shall use is that real analytic functions in a connected domain can be recovered from their derivatives at an internal point . ) for any @xmath81 and each integer @xmath80 , define the measure @xmath82 by @xmath83 \
, , \hspace{1cm}b\in{\mathcal b}_{\bar \omega_<^k}\ \,\ ] ] ( here @xmath84).an alternative description of @xmath85 is the following one : pick a random partition @xmath10 according to @xmath86 and then sample size - biased independently ( with replacement ) @xmath80 parts @xmath87 from @xmath10 . then , @xmath88 part of the proof of part ( b ) of theorem [ theo - uniq ] below will consist in verifying that these measures @xmath89 characterize @xmath86 ( see ( * ? ? ?
4 ) for a similar argument in a closely related context ) .
set for @xmath90 , @xmath91 our main result is part ( b ) of the following : * ( poisson - dirichlet law ) * [ theo - uniq ] assume @xmath41 $ ] and fix @xmath92 .
+ ( a ) the poisson - dirichlet law of parameter @xmath93 belongs to @xmath94 , and is invariant ( in fact : reversing ) for the kernel @xmath40 . + ( b ) assume a probability measure @xmath95 is @xmath43-invariant .
then @xmath86 is the poisson - dirichlet law of parameter @xmath93 .
the structure of the paper is as follows : in section [ sec - pre ] , we prove the feller property of @xmath40 , the existence of invariant probability measures for it , and some of their basic properties .
section [ sec - support ] and [ sec - trans ] are devoted to the proofs of theorems [ theo - duo ] and [ rectrans ] respectively , section [ sec - pois ] studies the poisson - dirichlet measures and provides the proof of theorem [ theo - uniq ] .
we conclude in section [ concluding ] with a list of comments and open problems .
for fixed @xmath35)$ ] , @xmath96 $ ] and @xmath97 we denote by @xmath98 the law of the markov process on @xmath99 with kernel @xmath43 and starting point @xmath10 , i.e. @xmath100=1 $ ]
. whenever @xmath101 , the law of the corresponding markov process with initial distribution @xmath86 is denoted by @xmath102 . in both cases , we use @xmath103 to denote the resulting process .
[ cont ] the kernel @xmath43 is feller , i.e.for any continuous function @xmath104 , the map @xmath105 @xmath106 is continuous .
we have @xmath107 one may assume that @xmath108 is of the form @xmath109 with @xmath110 and @xmath111 , since any @xmath112 can be uniformly approximated by such functions , and denote accordingly @xmath113 the @xmath114 norm of @xmath10 s first @xmath80 components .
we shall prove the lemma in this case by showing that both sums in ( [ kf ] ) contain finitely many nonzero terms , this number being uniformly bounded on some open neighborhood of a given @xmath115 , and that @xmath116 and @xmath117 are continuous for every @xmath118 .
for the second sum these two facts are trivial : @xmath119 and @xmath10 coincide in their first @xmath80 components @xmath120,\ \forall i > k$ ] , since splitting a component does nt affect the ordering of the larger ones , and thus @xmath121 for @xmath122 .
moreover , @xmath117 s continuity follows from equicontinuity of @xmath123 .
as for the first sum , given @xmath124 with positive components ( the necessary modification when @xmath115 has zero components is straightforward ) , let @xmath125 be such that @xmath126 and consider @xmath115 s open neighborhood @xmath127 . in particular
, for all @xmath128 , @xmath129 and thus , when @xmath130 , @xmath131 , which means that @xmath132 and @xmath10 coincide in their first @xmath80 components , or that @xmath133 for every @xmath134 and @xmath135 .
finally , each @xmath116 is continuous because the series defining it converges uniformly . indeed , for @xmath136 and uniformly in @xmath137 .
for a given @xmath138 , choose @xmath139 such that @xmath140 whenever @xmath141 .
then @xmath142 which proves the uniform convergence . [ schteim ]
let @xmath101 be @xmath43-invariant .
then @xmath143 furthermore , if we set for @xmath144 , @xmath145 then for all @xmath144 , @xmath146 let @xmath147 $ ] , and consider the random variable @xmath148 on @xmath149 which counts the intervals longer than @xmath150 .
we first prove ( [ half ] ) .
( the value @xmath151 is used in the subsequent proof of ( [ deu ] ) . )
assume that @xmath152 is finite which is always the case for @xmath153 since on @xmath149 , @xmath154 and is also true for @xmath151 if only finitely many @xmath155 are non zero .
then the expected ( conditioned on @xmath10 ) increment @xmath156 of @xmath152 after one step of the underlying markov process is well - defined .
it equals @xmath157- { \sigma}[[1-{\varepsilon}/p_i,1/2]]\right)\nonumber\\ & & - \beta_m \sum_i p_i^2 \left(1_{p_i\leq { \varepsilon}<2p_i}- 1_{{\varepsilon}<p_i}\right).\nonumber\end{aligned}\ ] ] the right hand side of ( [ flug ] ) converges as @xmath150 tends to 0 to @xmath158 since @xmath86 is @xmath43-invariant we have @xmath159 for all @xmath150 .
now ( [ half ] ) follows from ( [ rishon ] ) by dominated convergence since @xmath160 . for the proof of ( [ deu ] ) note that for all @xmath161 , @xmath162 has full measure on sequences @xmath163 for which the number @xmath164 of nonvanishing components is finite because we start with @xmath165 @xmath166-a.s . and
@xmath164 can increase at most by one in each step .
given such a @xmath163 , the expected increment @xmath167 of @xmath164 equals ( see ( [ flug ] ) , ( [ rishon ] ) ) @xmath168 .
therefore for @xmath169 , @xmath170 summing over @xmath171 yields @xmath172 the left hand side of ( [ liba ] ) is nonnegative due to @xmath173 and @xmath174 .
this proves ( [ deu ] ) .
[ exists ] there exists a @xmath43-invariant probability measure @xmath44 .
define @xmath162 and @xmath175 as in ( [ domi ] ) .
since @xmath149 is compact , @xmath176 is compact .
consequently , there are @xmath177 and a strictly increasing sequence @xmath178 of positive integers such that @xmath179 converges weakly towards @xmath86 as @xmath180 .
this limiting measure @xmath86 is invariant under @xmath43 by the following standard argument . for any continuous function @xmath104 , @xmath181\\ & = & \lim_{n\to\infty}\frac{1}{m_n}\sum_{k=0}^{m_n-1 } \int\int f\
dk_{{\sigma},\beta_m,\beta_s}(p,\cdot)\ d\nu_k(p)\\ & = & \lim_{n\to\infty}\frac{1}{m_n}\sum_{k=0}^{m_n-1 } \int f(p)\ d\nu_{k+1}(p ) = \lim_{n\to\infty } \int f d\bar{\nu}_{m_n}=\int f d\mu\,.\end{aligned}\ ] ] hence it remains to show that @xmath4 has full @xmath86-measure , i.e. @xmath182=1 $ ] . to prove this observe that @xmath183 ( unlike @xmath184 ) is a continuous function on @xmath185
therefore by ( [ half ] ) , weak convergence and ( [ deu ] ) , @xmath186 by which the first inequality is an equality , and thus @xmath187 [ prop ] if @xmath44 is @xmath188-invariant for @xmath189 , then @xmath190 and @xmath191 .
let @xmath192 be an integer and @xmath193 .
given @xmath10 , consider the expected increment @xmath194 of @xmath195 after one step of the process driven by @xmath196 : @xmath197 note that @xmath198 is finite because of @xmath199 .
therefore , by invariance , @xmath200 , which implies @xmath201=\frac { \beta_{m,\alpha}\int \sum_{i\ne j}p_i p_j \left(p_i^k+p_j^k-(p_i+p_j)^k\right)\ d\mu } { \int \sum_ip_i^{2+k}\ d\mu}.\ ] ] hence , for any @xmath80 , @xmath202 taking @xmath203 we conclude that @xmath204 .
this proves the second claim .
in addition , we have @xmath205 for all @xmath192 .
obviously , ( [ shalom ] ) also holds true for @xmath206 .
extend @xmath207 to probability measures on @xmath0 $ ] which are supported on @xmath208 $ ] .
it is enough for the proof of @xmath190 to show that for all continuous real valued functions @xmath209 on @xmath0 $ ] which vanish on @xmath210 $ ] the integrals @xmath211 coincide for @xmath212 .
fix such an @xmath209 and choose a sequence of polynomials @xmath213 which converges uniformly on @xmath0 $ ] to @xmath209 as @xmath180 .
then @xmath214 converges uniformly on @xmath0 $ ] to @xmath215 .
since @xmath216 vanishes on the support of @xmath217 and @xmath218 we get for @xmath219 , @xmath220 which is the same for @xmath221 and @xmath222 due to ( [ shalom ] ) .
theorem [ theo - duo ] is a consequence of the following result .
[ bit ] let @xmath223 be @xmath40-invariant and denote @xmath224 and @xmath225 s stopping time @xmath226 .
then @xmath227=1 \longleftrightarrow \mu[s]>0 \longleftrightarrow \mu[\{\bar{p}\}]>0 \longleftrightarrow e_{\bar{p}}[h]<\infty.\ ] ] we start by proving that @xmath228 implies @xmath229=1 $ ] . fix an arbitrary @xmath230 and consider the random variables @xmath231 after one step of the process @xmath232 may increase , decrease or stay unchanged . if we merge two intervals then @xmath232 can not decrease , but may increase by the mass of one or two intervals which are smaller than @xmath233 but become part of an interval which is bigger than @xmath233 .
if we split an interval then @xmath232 can not increase , but it decreases if the original interval was larger than @xmath233 and at least one of its parts is smaller than @xmath233 . thus given @xmath10 , the expected increment @xmath234 of @xmath232 after one step of the process is @xmath235 we bound @xmath236 from below by @xmath237 where @xmath238 and @xmath239 from above by @xmath240}\\ & & + \
\beta_s \sum_{i\geq 1}\int \sum_{j=0}^{n-1}p_i^3 x 1_{{\vartheta}^{n - j}<p_i\leq { \vartheta}^{n - j-1 } } 1_{p_i\leq { \vartheta}^n / x}\ d{\sigma}(x)\\ & \leq & \beta_s \sum_{i\geq 1}{\vartheta}^{3(n-1 ) } 1_{{\vartheta}^n < p_i\leq { \vartheta}^{n-1}}\\ & & + \beta_s\ \sum_{i\geq 1}\int \sum_{j=0}^{n-1}{\vartheta}^{3(n - j-1 ) } x 1_{{\vartheta}^{n - j}<p_i\leq { \vartheta}^{n - j-1 } } 1_{x\leq { \vartheta}^j}\ d{\sigma}(x)\\ & \leq&\beta_s { \vartheta}^{3(n-1 ) } \sharp i_n + \beta_s \
\sum_{j=0}^{n-1}\sum_{i\geq 1}{\vartheta}^{3(n - j-1 ) } { \vartheta}^j 1_{{\vartheta}^{n - j}<p_i\leq { \vartheta}^{n - j-1 } } { \sigma}[(0,{\vartheta}^j]]\\ & \leq&\beta_s { \vartheta}^{3(n-1 ) } \sharp i_n + \beta_s { \vartheta}^{3(n-1)}\sum_{j=0}^{n-1 } { \vartheta}^{-2j}{\sigma}[(0,{\vartheta}^j]]\sharp i_{n - j}\\ & \leq&2 \beta_s { \vartheta}^{3(n-1)}\sum_{j=0}^{n-1}{\vartheta}^{-2j}{\sigma}[(0,{\vartheta}^j ] ] \sharp i_{n - j}.\end{aligned}\ ] ] since @xmath86 is invariant by assumption , @xmath241 and therefore @xmath242 \int \sharp i_{n - j}\ d\mu\\ & = & 2 \beta_s { \vartheta}^{-5}\sum_{j=0}^{n-1}{\vartheta}^{-j}{\sigma}[(0,{\vartheta}^j]]\int { \vartheta}^{n - j}\sharp i_{n - j}\ d\mu.\end{aligned}\ ] ] consequently , @xmath243}\\ & \leq & \sum_{n\geq 1}2\int w_n \sharp i_{n+1}\ d\mu \ \leq\ \frac{2{\vartheta}^{-5}\beta_s}{\beta_m } \sum_{n\geq 1}\sum_{j=0}^{n-1}{\vartheta}^{-j}{\sigma}[(0,{\vartheta}^j ] ] \int { \vartheta}^{n - j}\sharp i_{n - j}\ d\mu\\ & = & \frac{2 { \vartheta}^{-5}\beta_s}{\beta_m } \left(\sum_{j=0}^\infty{\vartheta}^{-j}{\sigma}[(0,{\vartheta}^j]]\right ) \sum_{n\geq 1 } \int { \vartheta}^{n}\sharp i_{n } \ d\mu\\ & \leq & \frac{2{\vartheta}^{-5}\beta_s}{(1-{\vartheta})\beta_m } \left(\sum_{j=0}^\infty({\vartheta}^{-j}-{\vartheta}^{-j+1 } ) { \sigma}[(0,{\vartheta}^j]]\right ) \sum_{n\geq 1 } \int \sum_{i\in i_n}p_i\ d\mu\\ & = & \frac{2{\vartheta}^{-5}\beta_s}{(1-{\vartheta})\beta_m } \left(\int\sum_{j=0}^\infty 1_{{\vartheta}^{-j}\leq 1/x } ( { \vartheta}^{-j}-{\vartheta}^{-j+1})\ d{\sigma}(x)\right ) \int |p|\ d\mu\\ & \leq & \frac{2{\vartheta}^{-5}\beta_s}{(1-{\vartheta})\beta_m}\int\frac{1}{x}\ d{\sigma}(x)\end{aligned}\ ] ] which is finite by assumption
. thus by borel - cantelli , @xmath244 is @xmath86-a.s .
eventually ( for large @xmath66 ) less than 1/2
. however , @xmath232 converges @xmath86-a.s . to 1 as @xmath66 tends to @xmath245 .
thus even @xmath246 is @xmath86-a.s .
eventually less than 1/2 , which means that @xmath247 is @xmath86-a.s .
eventually empty , that is @xmath229=1 $ ] .
now we assume @xmath229>0 $ ] in which case there exist some @xmath248 and @xmath153 such that @xmath249>0.$ ] by @xmath118 successive merges of the positive parts and @xmath86 s invariance we obtain @xmath250=\mu[p_1=1]\geq ( 2\beta_m { \varepsilon}^2)^{i-1}\delta > 0.\ ] ] next , we assume @xmath251>0 $ ] and note that @xmath252 and thus , defining @xmath253 $ ] , one obtains an invariant measure supported on @xmath254 .
the chain determined by @xmath43 on @xmath254 is @xmath255-irreducible , and has @xmath256 as invariant measure , with @xmath257>0 $ ] .
therefore , kac s recurrence theorem ( * ? ? ?
* theorem 10.2.2 ) yields @xmath258<\infty$ ] .
finally , we assume @xmath258<\infty$ ] and show @xmath48 .
if @xmath259 , then @xmath260=\beta_s>0 $ ] , and when @xmath261 we write @xmath262 , where @xmath263 has distribution @xmath3 . furthermore ,
restricted to @xmath63 and conditioned on @xmath263 , @xmath264a.s .
, where in terms of the chain s sampling and merge / split interpretation , @xmath265 is the first time a marked part of size @xmath263 is sampled , i.e. a geometric random variable with parameter @xmath266 .
thus @xmath267\ge p_{\overline{p}}[a]e_{\overline{p}}[h|a ] \ge \beta_s\left(1+\int e_{p^\xi}[\tau]d\sigma(\xi)\right ) \ge \beta_s\left(1+\int \frac{1}{2\xi}d\sigma(\xi)\right).\ ] ] [ coro ] if @xmath268 then there exists a unique @xmath43-invariant probability measure @xmath44 . in view of theorem
[ theo - duo ] , for the study of invariant measures it is enough to restrict attention to the state space @xmath254 , where the markov chain @xmath269 is @xmath255-irreducible , implying , see ( * ? ? ?
* chapter 10 ) , the uniqueness of the invariant measure .
the statement about positive recurrence is included in theorem [ bit ] .
the idea for the proof of the transience statement is to show that under ( [ bush ] ) the event that the size of the smallest positive part of the partition never increases has positive probability .
by @xmath270 we enumerate the times @xmath271 at which the value of the markov chain changes .
denote by @xmath272 the ( random ) number of instants among the first @xmath66 steps of the markov chain in which some interval is split . since @xmath273 is the number of steps among the first @xmath271 steps in which two parts are merged and since this number can never exceed @xmath274 if @xmath275 , we have that @xmath276-a.s .
, @xmath277 let @xmath278 denote the times at which some part is split .
this part is split into two parts of sizes @xmath279 and @xmath280 with @xmath281 .
according to the model the random variables @xmath282 are independent with common distribution @xmath283 .
further , for any deterministic sequence @xmath284 , let @xmath285 $ ] denote the law of the process which evolves using the kernel @xmath40 except that at the times @xmath286 it uses the values @xmath287 as the splitting variables .
note that @xmath288=\int p_{\xi,\bar p}[\ \cdot\ ] \
d\sigma^{{{\mathbb{n}}}}(\xi)\,.\ ] ] now denote by @xmath289 the size of the smallest positive part at time @xmath66 .
we prove that for @xmath290 , @xmath291 ( here and in the sequel , we take @xmath292 if @xmath293 ) .
indeed , we need only consider the case @xmath294 , in which case there exists a @xmath295 such that @xmath296 , and @xmath297 . but
clearly @xmath298 , and then the condition @xmath299 and the fact that @xmath297 imply @xmath300 , as claimed .
next , fix some @xmath301 $ ] where @xmath302 .
we will prove by induction over @xmath303 that @xmath304\geq \beta_s 1_{\xi_1<{\varepsilon}}\prod_{k=1}^{j-1}\left(1-\frac{\xi_1\wedge\ldots\wedge \xi_{\lceil k/2\rceil}}{\beta_0}\right).\ ] ] for @xmath305 the left hand side of ( [ hnpp ] ) equals the probability that the unit interval is split in the first step with the smaller part being smaller than @xmath150 which equals @xmath306 .
assume that ( [ hnpp ] ) has been proved up to @xmath8 .
then , with @xmath307 , @xmath308}\\ & = & e_{\xi,\bar{p}}\left [ p_{\xi,\bar{p}}[q(n_j)\geq q(n_{j+1})\ |\ { \mathcal f}_{n_j } ] , { \varepsilon}>q(1),\ q(0)\geq \ldots\geq q(n_j)\right].\label{wer}\end{aligned}\ ] ] now choose @xmath80 minimal such that @xmath309 .
one possibility to achieve @xmath310 is not to merge the part @xmath311 in the next step in which the markov chain moves .
the probability to do this is @xmath312 therefore ( [ wer ] ) is greater than or equal to @xmath313.\ ] ] by ( [ impl ] ) this can be estimated from below by @xmath314.\ ] ] this is due to ( [ as ] ) greater than or equal to @xmath315.\ ] ] along with the induction hypothesis this implies ( [ hnpp ] ) for @xmath316 .
taking expectations with respect to @xmath263 in ( [ hnpp ] ) yields @xmath317 \geq e_{\bar{p}}\left[\beta_s1_{\xi_1<{\varepsilon}}\prod_{k\geq 1}\left(1-\frac{{\varepsilon}\wedge\xi_2\wedge\ldots\wedge \xi_{\lceil k/2\rceil}}{\beta_0}\right)\right].\ ] ] by independence of @xmath318 from @xmath319 the right hand side of ( [ warum ] ) equals @xmath320 e_{\bar{p}}\left[\prod_{k\geq 2}\left(1-\frac{{\varepsilon}\wedge\xi_2\wedge\ldots\wedge \xi_k}{\beta_0}\right)^2\right].\ ] ] observe that ( [ bush ] ) implies @xmath321={\sigma}[(0,{\varepsilon})]>0 $ ] . by jensen s inequality and monotone convergence , ( [ stoned ] ) can be estimated from below by @xmath322\right)\ ] ] with some positive constant @xmath323 . since @xmath324 for @xmath325 $ ] this is greater than @xmath326\right)=c_1\exp\left(-\frac{4}{\beta_0 } \int_0^{1/2 } \frac{p_{\bar{p}}[\xi_1 > t]}{p_{\bar{p}}[\xi_1\leq t]}\ dt\right)\ ] ] where we used that due to independence @xmath327=\int_0^{1/2}p_{\bar{p}}[\xi_1
> t]^{k-1}\ dt.\ ] ] due to assumption ( [ bush ] ) , ( [ oje ] ) and therefore also the left hand side of ( [ warum ] ) are positive .
this implies transience of @xmath50 .
throughout this section , the splitting measure is the uniform measure on @xmath36 $ ] . to emphasize this
, we use @xmath328 instead of @xmath40 throughout . recall that @xmath92
. it will be convenient to equip @xmath57 ( but not @xmath25 ) with the @xmath18 topology ( noting that the borel @xmath3-algebra is not affected by this change of topology ) , and to replace the kernel @xmath329 by @xmath330 both kernels coincide on @xmath4 ( not on @xmath331 ) .
however , @xmath332 has the advantage that it is well defined on all of @xmath57 and is homogeneous ( hence the superscript @xmath333 ) in the sense of the first of the following two lemmas , whose proof is straightforward and in which by a slight abuse of notation @xmath332 will denote both the kernel in @xmath4 and in @xmath57 and also the operators induced by these kernels , the distinction being clear from the context .
[ homogeneous ] for all @xmath334 , @xmath335 .
+ more generally , denoting @xmath336 , we have @xmath337 .
+ in particular , if @xmath338 is invariant ( resp . reversing ) for @xmath332
then + @xmath339 is invariant ( resp . reversing ) for @xmath329 .
[ feller ] the kernel @xmath332 maps continuous bounded functions to continuous bounded functions . note that we work with the @xmath18 topology , and hence have to modify the proof in lemma [ cont ] .
the @xmath18 topology makes the mapping @xmath340 continuous ( when @xmath4 is equipped with the induced @xmath18 topology ) .
fix @xmath341 . by ( [ kh ] )
we have @xmath342 note that for @xmath343 is of the form @xmath344 , with @xmath345 , and @xmath346 denoting the standard duality pairing . in stating this we have used the facts that @xmath347 is continuous and bounded , and that all the mappings @xmath33 and @xmath34 are contractive .
the continuity of @xmath348 , then follows from @xmath349 after observing that @xmath350 and @xmath351 remain bounded in any @xmath18 neighborhoods of @xmath10 .
the continuity of @xmath352 is obvious being the product of two continuous functions of @xmath10 .
it has thus been shown that @xmath353 .
[ reversible ] the poisson - dirichlet measure @xmath354 is reversing for @xmath355 with @xmath356 $ ] . by lemma [ homogeneous ]
it suffices to verify that @xmath357 is reversing for the kernel @xmath332 , which for simplicity will be denoted by @xmath358 for the rest of this proof .
we thus need to show that @xmath359 because @xmath360 and @xmath361 are absolutely continuous with respect to @xmath58 , it follows from ( [ khf ] ) that if @xmath362 are uniformly bounded functions such that @xmath363 , then @xmath364 .
thus , by standard density arguments we may and shall assume @xmath365 and @xmath366 to be continuous .
define for each @xmath153 the truncated intensity measure @xmath367 , and the corresponding poisson measure @xmath368 , with expectation operator @xmath369 .
alternatively , if @xmath26 is distributed in @xmath57 according the @xmath58 , then @xmath368 is the distribution of @xmath370 , that is , @xmath371 . observe that @xmath372 , @xmath373 implying that the measures @xmath368 converge weakly to @xmath58 as @xmath374 . to prove ( [ gkf = fkg ] ) we first write @xmath375 and
conclude that the first and third terms in ( [ three ] ) converge to @xmath376 as @xmath374 by virtue of the weak convergence of @xmath377 to @xmath58 and @xmath358 s feller property , established in lemma [ feller ] .
it thus remains to be shown that , for all @xmath378 and @xmath153 , @xmath379 the truncated intensity @xmath380 has finite mass @xmath381 , and thus @xmath382 , @xmath368a.s .
in particular each @xmath383 can be naturally represented as a sequence @xmath384 of symmetric @xmath385 s @xmath386 , with @xmath387 for each @xmath66 . as a result , and in terms of the expectation operators @xmath388 of @xmath368 conditioned on @xmath389 , we may write @xmath390,\ ] ] while by the definition ( [ kh ] ) of @xmath332 and the properties stated above of the poisson random measure conditioned on @xmath389 , @xmath391 where @xmath392 .
our goal is to prove that this expression , after summing in @xmath66 , is roughly symmetric in @xmath365 and @xmath366 ( as stated precisely in ( [ approxrevers ] ) ) .
obviously @xmath393 , and in addition we aim at showing that @xmath394 ( with an error appropriately small as @xmath374 ) .
this will be achieved by a simple change of variables , including the splitting coordinate @xmath395 in @xmath396 . in the integral of the @xmath118th term in @xmath397
perform the change of variables @xmath398 given by @xmath399 ( or @xmath400 ) .
more precisely , @xmath401 + for which @xmath402 and @xmath403 , so that + @xmath404 @xmath405 ( @xmath406 is as the term preceding it but with the @xmath407 and @xmath408 integrals taken in @xmath409 $ ] , and the notation @xmath410 means that the variable @xmath411 has been eliminated from the denominator ) @xmath412 ( by @xmath413 s symmetry , the sum s @xmath414 terms are equal , hence the last equality ) .
+ on the other hand , and for the same reason of symmetry , the @xmath415 terms in @xmath416 are all equal so that @xmath417 comparing ( [ i2 ] ) with ( [ i1 ] ) , and observing that by definition @xmath418 , we conclude that there exists a @xmath419 such that , for @xmath420 , @xmath421 applying ( [ cenbound ] ) via ( [ ekfg ] ) in ( [ fg - gf ] ) twice , once as written and once reversing the roles of @xmath365 and @xmath366 , and noting that @xmath422 , we have @xmath423 from which ( [ approxrevers ] ) follows immediately since @xmath424 .
\(a ) the poisson - dirichlet law @xmath425 is reversing by theorem [ reversible ] , and hence invariant .
we now show that it belongs to @xmath73 .
note first that @xmath85 is absolutely continuous with respect to @xmath79 : for any @xmath426 with @xmath427 , it holds that @xmath428 d\gamma_\theta(x)=0\,,\ ] ] where we used the fact that under @xmath58 , @xmath429 and @xmath430 are independent , with @xmath430 being distributed according to the gamma law @xmath431 of density @xmath432 ( see @xcite ) .
it thus suffices to compute the limit @xmath433}{\delta^k}\,,\ ] ] where all @xmath434 are pairwise distinct and nonzero , to have @xmath435 for such @xmath436 , set @xmath437 and @xmath438 .
define @xmath439 by the memoryless property of the poisson process , for any borel subset @xmath440 , @xmath441 where ( [ limp ] ) , as above , is due to @xcite .
further , recall that @xmath442 and @xmath443 are independent .
recall that the density of the poisson process at @xmath444 is @xmath445 , where @xmath446 . performing the change of variables @xmath447
, one finds that the jacobian of this change of coordinate is @xmath448 ( in computing this jacobian , it is useful to first make the change of coordinates @xmath449 where @xmath450 are considered as independent coordinates , and note the block - diagonal structure of the jacobian ) .
it follows that @xmath451 which is real analytic on @xmath452 .
thus , @xmath453 . in passing , we note that @xmath454 on @xmath78 when @xmath71 .
\(b ) 1 ) first we show that the family of functions @xmath455 associated with @xmath86 , determines @xmath86 . to this end , define for @xmath456 functions @xmath457 $ ] by @xmath458 note that any function @xmath459 with @xmath460 can be written after expansion of the product as a ( finite ) linear combination of functions @xmath461 with @xmath462 . since we have by the definition of @xmath85 that @xmath463 the family @xmath455 therefore determines the expectations @xmath464 .
consequently , @xmath455 determines also the joint laws of the random variables @xmath465 under @xmath86 .
we claim that these laws characterize @xmath86 .
indeed , let @xmath256 be the distribution of the random variable @xmath466^{{\mathbb{n}}}$ ] under @xmath86 . since @xmath467 is injective it suffices to show that the distributions of @xmath465 under @xmath86 determine @xmath256 .
but , since any continuous test function @xmath365 on the compact space @xmath0^{{\mathbb{n}}}$ ] can be uniformly approximated by the local function @xmath468 , this is true due to @xmath469 \2 ) for @xmath95 , the set of numbers @xmath470 with @xmath192 and @xmath471 yare enough to characterize @xmath472 , and hence by the first part of the proof of b ) , to characterize @xmath86 .
it is thus enough to prove that @xmath328 uniquely determines these numbers . toward this end ,
first note that @xmath473=1.\ ] ] to simplify notations , we define @xmath474 and extend @xmath475 to a function on @xmath0^k$ ] by setting it 0 on the complement of @xmath78 .
for @xmath192 we have @xmath476 indeed , for @xmath477 this is ( [ heat ] ) while for @xmath478 , and arbitrary @xmath479 , @xmath480\times b]}\\ & = & e_\mu \bigg[\sum_{(j_2,\ldots , j_k)\in\mathbb{n}^{k-1}_{\neq } } \left(\prod_{i=2}^kp_{j_i}\right)\,{\mathbf 1}_{b}(p_{j_2},\ldots , p_{j_k})\sum_{j_1\notin \{j_2,\ldots , j_k\}}p_{j_1}1_{[0,1]}(p_{j_1})\bigg]\\ & = & e_{\mu_{k-1}}\left[1_b(p_2,\ldots , p_k)\left(1-\sum_{i=2}^k p_i\right)\right]\\ & = & \int_b \left(1-\sum_{i=2}^k x_i\right)m_{k-1}(x_2,\ldots , x_k)\ dx_2\ldots dx_k,\end{aligned}\ ] ] which implies ( [ fifth ] ) .
now we fix @xmath192 , apply @xmath328 to the test function @xmath481 , with @xmath482 having pairwise distinct coordinates , and take @xmath483 , which yields the basic relation @xmath484 here the left hand side represents mergings and splittings that produce a new part roughly at one of the @xmath434-s ; the right hand side represents parts near one of the @xmath434-s that merge or split . after multiplying by @xmath485 , rearranging and using ( [ fifth ] ) to get rid of the integral with upper limit 1
, we obtain the equality @xmath486 we now proceed to show how ( [ frogs ] ) ( [ second ] ) yield all the required information . as starting point for a recursion
, we show how to compute @xmath487 for all @xmath192 . taking in ( [ frogs ] ) ( [ second ] ) all @xmath488 except for @xmath489 and using the continuity of the functions @xmath475 yields @xmath490 letting @xmath491 we get @xmath492 . with @xmath493 as start of the recursion this implies @xmath494 for the evaluation of the derivatives of @xmath475 we proceed inductively .
recall the functions @xmath495 defined in ( [ corec ] ) , and write @xmath496 , @xmath497 , for @xmath498 .
fix @xmath499 such that @xmath500 , with @xmath501 .
our analysis rests upon differentiating ( [ frogs ] ) ( [ second ] ) @xmath502 times with respect to @xmath489 ; to make this differentiation easy , call a term a _
g term of degree @xmath503 _ if it is a linear combination of terms of the form @xmath504 and @xmath505 and @xmath506 and @xmath507 note that ( [ frogs ] ) ( [ second ] ) contains one g term of degree @xmath508 in ( [ frogs ] ) and that differentiating a g term of degree @xmath503 once yields a g term of degree @xmath509 . thus , differentiating the g term in ( [ frogs ] ) @xmath501 times and substituting @xmath510 , we recover a constant multiple of @xmath511 . similarly , call a term an _
h term of degree @xmath503 _ if it is a linear combination of terms of the form @xmath512 observe , that differentiating an h term of degree @xmath503 produces an h term of degree @xmath509 .
if we differentiate twice the term @xmath513 in ( [ cows ] ) we get an h term of degree 0 . therefore differentiating this term @xmath501 times results in an h term of degree @xmath514 .
since the term @xmath515 in ( [ second ] ) is an h term of degree -2 , differentiating this term @xmath502 times produces also an h term of degree @xmath514 .
thus both terms produce after @xmath502-fold differentiation and evaluation at @xmath510 a constant multiple of @xmath516 .
the h term @xmath517 in ( [ second ] ) is treated more carefully .
it is easy to see by induction that its @xmath502-th derivative equals @xmath518 .
evaluating it at @xmath510 gives @xmath519 .
moreover , the terms in ( [ magrefa ] ) and ( [ united ] ) for @xmath520 vanish when differentiated twice with respect to @xmath489 . summarizing the above
, we conclude by differentiating ( [ frogs ] ) ( [ second ] ) @xmath501 times with respect to @xmath489 and subsequent evaluation at @xmath510 that there are some constants @xmath521 , such that @xmath522}\\ & = & \nonumber c_1 m_{k+1}^{(n_1 - 2)}(0,x_2,\ldots , x_k,0)\qquad[(\ref{frogs}),i=1]\\ & & + c_2 m_k^{(n_1 - 2)}(0,x_2,\ldots , x_k)\qquad[(\ref{second}b),i=1 + ( \ref{cows}),i=1]\\ & & \label{lauf } -\left[\beta_m \left(\sum_{i=2}^k x_i\right)+(\beta_m-\beta_s ) \left(\sum_{i=2}^k x_i^2\right)\right ] m_k^{(n_1)}(0,x_2,\ldots , x_k)\qquad[(\ref{second})]\\ & & + \nonumber \beta_m \sum_{i=2}^k x_i \int_0^{x_i } m_{k+1}^{(n_1)}(0,x_2,\ldots , x_{i-1},z , x_i - z , x_{i+1},\ldots , x_k)dz \qquad[(\ref{frogs})]\\ & & - \beta_s \sum_{i=2}^k x_i \int_0^{x_i } m_{k}^{(n_1)}(0,x_2,\ldots , x_{i-1},z , x_{i+1},\ldots , x_k)dz \qquad[(\ref{cows})]\nonumber \\ & & + \beta_s\sum_{i=2}^k x_i\nonumber m_{k-1}^{(n_1)}(0,x_2,\ldots , x_{i-1},x_{i+1},\ldots , x_k)\qquad[(\ref{magrefa})]\\ & & -\beta_s n_1 \sum_{i=2}^k x_i
m_{k-1}^{(n_1 - 1)}(0,x_2,\ldots , x_{i-1},x_{i+1},\ldots , x_k ) \qquad[(\ref{united } ) , j=1]\nonumber\\ & & -\beta_s \sum_{i=2}^k\sum_{j=2 , j\ne i}^k x_i x_j m_{k-1}^{(n_1)}(x_1,\ldots , x_{i-1},x_{i+1},\ldots , x_k)\qquad[(\ref{united})]\end{aligned}\ ] ] for @xmath523 only the first three lines do not vanish and give a recursion which allows us to compute starting with ( [ hit ] ) all derivatives @xmath524 .
further differentiating with respect to @xmath525 , one concludes that there exist constants @xmath526 such that @xmath527 \nonumber \\ & + & \!\!\!\!\!\!\ ! \!\!\!\!\!\!\ ! \!\!\!\!\!\!\ ! \sum_{{{\bf n}}':|{{\bf
n}}'|\leq |{{\bf n}}|-1 , n_i'\leq n_i , n_1=n_1 ' } [ d^4_{{{\bf n}},{{\bf n } } ' } m_k^{({{\bf n}}')}+d^5_{{{\bf n}},{{\bf n } } ' } m_{k+1}^{({{\bf n}}',0 ) } + d^6_{{{\bf n}},{{\bf n } } ' } m_{k-1}^{({{\bf n}}')}]\,,\end{aligned}\ ] ] where @xmath528 unless @xmath529 possesses at least one component which is zero .
we now compute iteratively any of the @xmath530 , with @xmath500 : first , substitute in ( [ shweishwei ] ) @xmath531 to compute @xmath532 , for all @xmath533
. then , substitute @xmath534 ( @xmath535 ) to compute iteratively @xmath536 from the knowledge of the family @xmath537 , etc .
more generally , having computed the terms @xmath538 , we compute first @xmath539 by substituting in ( [ shweishwei ] ) @xmath540 , and then proceed inductively as above .
\1 ) we of course conjecture the it is tempting to use the technique leading to ( [ half ] ) in order to prove the conjecture by characterizing the expectations with respect to @xmath86 of suitable test functions .
one possible way to do that is to consider a family of polynomials defined as follows .
let @xmath541 be a finite sequence of nonnegative integers , with @xmath542 .
we set @xmath543 , i.e. we consider @xmath544 as representing a partition of @xmath545 having @xmath271 parts of size @xmath8 , and no parts of size @xmath5 .
recall next @xmath546 and the * n*-_polynomial _
@xmath547 @xmath545 is the _ degree _ of @xmath548 , and , with @xmath499 and @xmath549 as above , @xmath549 is the _ maximal monomial degree _ of @xmath550 . because we do not allow partitions with parts of size @xmath5
, it holds that @xmath551 if @xmath552 ( i.e , there exists a point @xmath7 such that @xmath553 ) .
it is easy to check that the family of polynomials @xmath554 is separating for @xmath555 . letting @xmath556 denote the expected increment ( conditioned on @xmath10 ) of @xmath550 after one step of the process
, we have that @xmath556 is uniformly bounded .
hence , by invariance of @xmath86 , @xmath557 . expanding this equality , we get that @xmath558= \nonumber\\ & & \!\!\!\!\ ! \!\!\!\!\ ! \!\!\!\!\ ! -e_\mu\left [ \sum_\alpha p_\alpha^2 \sum_{k=2}^d \int\!\!\!\left(\!\left(\prod_{j=2}^{k-1 } \!\left(z^f_{j,\alpha , x}\right)^{n_j}\right ) \!\!\left(\sum_{\ell=0}^{n_k-1 } z_k^\ell\ ! \left(\begin{array}{l }
\!\!n_k\\ \!\!\ell \!\!\end{array}\!\!\right)\ !
f_{\alpha , k , x}^{n_k-\ell}\right ) \!\left(\prod_{j = k+1}^d z_j^{n_j}\right ) \!\right)\!d\sigma(x)\right ] \nonumber \\ & & \!\!\!\!\!\!\!\!\!\!\ !
+ \frac{\beta_m}{\beta_s } e_\mu \left[\sum_\alpha p_\alpha^2 \sum_{k=2}^d\left(\prod_{j=2}^{k-1 } \left(z_j+(2^j-2)p_\alpha^j\right)^{n_j}\right ) \right .
\nonumber\\ & & \;\;\;\;\;\;\;\;\;\;\ ; \;\;\;\;\;\;\;\;\;\;\ ; \;\;\;\;\;\;\;\;\;\;\ ;
\left.\left(\sum_{\ell=0}^{n_k-1}z_k^\ell \left(\begin{array}{l } \!\!n_k\\ \!\!\ell \end{array}\!\!\right ) \left ( ( 2^k-2)p_\alpha^k\right)^{n_k-\ell}\right ) \left(\prod_{j = k+1}^d z_j^{n_j}\right)\right]\end{aligned}\ ] ] where @xmath559p_\alpha^j\leq 0\,,\ ] ] @xmath560 note that all terms in ( [ aroch ] ) are positive .
note also that the right hand side of ( [ aroch ] ) is a polynomial of degree @xmath561 , with maximal monomial degree @xmath562 , whereas the left hand side is a polynomial of degree at most @xmath563 and maximal monomial degree at most @xmath549 .
let @xmath564 denote the number of integer partitions of @xmath80 which do not have parts of size @xmath5 .
then , there are @xmath564 distinct polynomials of degree @xmath80 , whereas ( [ aroch ] ) provides at most @xmath565 relations between their expected values ( involving possibly the expected value of lower order polynomials ) .
since always @xmath566 , it does not seem possible to characterize an invariant probability measure @xmath567 using only these algebraic relations .
\4 ) there is much literature concerning coagulation - fragmentation processes .
most of the recent probabilistic literature deals with processes which exhibit either pure fragmentation or pure coagulation . for an extensive review , see @xcite , and a sample of more recent references is @xcite , @xcite and @xcite .
some recent results on coagulation - fragmentation processes are contained in @xcite . however , the starting point for this and previous studies are the coagulation - fragmentation equations , and it is not clear how to relate those to our model .
the functions @xmath475 introduced in the context of theorem [ theo - uniq ] are related to these equations .
\5 ) a characterization of the poisson - dirichlet process as the unique measure coming from an i.i.d .
residual allocation model which is invariant under a _ split and merge _ transformation is given in @xcite .
j. pitman has pointed out to us that a slight modification of this transformation , preceded by a size biased permutation and followed by ranking , is equivalent to our markov transition @xmath40 .
pitman @xcite then used this observation to give an alternative proof of part ( a ) of theorem [ theo - uniq ] .
\6 ) yet another proof of part a ) of theorem [ theo - uniq ] which avoids the poisson representation and theorem [ reversible ] can be obtained by computing the expectation of the polynomials @xmath570 , defined in remark 1 ) above , under the poisson - dirichlet law .
we prefer the current proof as it yields more information and is more transparent .
\7 ) a natural extension of poisson - dirichlet measures are the _ two parameter _ poisson - dirichlet measures , see e.g. @xcite .
pitman raised the question , which we have not addressed , of whether there are splitting measures @xmath3 which would lead to invariant measures from this family .
\8 ) while according to theorem [ theo - uniq ] there is a reversing probability measure for @xmath356 $ ] this does not hold for general @xmath35)$ ] .
for instance , let us assume that the support of @xmath283 is finite .
then there exist @xmath571 such that @xmath572=0 $ ] . to show that any invariant measure @xmath86 is not reversing it suffices to find @xmath573 such that the detailed balance equation @xmath574 k_{{\sigma},\beta_m,\beta_s}(s,\{t\})= \mu[\{t\ } ] k_{{\sigma},\beta_m,\beta_s}(t,\{s\})\ ] ] fails . due to theorem [ bit ] ,
@xmath251>0 $ ] .
now we first refine the partition @xmath50 by successive splits until we reach a state @xmath163 with @xmath575 , where @xmath153 is a small number . since @xmath86 has finite support , @xmath576>0 $ ] .
then we create from @xmath10 by successive mergings some @xmath577 with @xmath578 , which is possible if @xmath150 was chosen small enough . again , @xmath579>0 $ ] .
if we call now @xmath580 the state which one gets from @xmath581 by merging @xmath582 and @xmath583 , then the left hand side of ( [ son ] ) is positive . on the other hand ,
the right hand side of ( [ son ] ) is zero because of @xmath584 due to the choice of @xmath585 and @xmath586 .
* acknowledgment * we thank r. brooks for suggesting to us the study of the fragmentation - coagulation process described here , a. vershik for the reference to @xcite and some useful discussions concerning the poisson - dirichlet measure , and a .- s .
sznitman for discussions concerning the uniqueness issue , and for his help with the proof of theorem [ reversible ] . finally , we thank j. pitman for pointing out @xcite , making his observation @xcite available to us , and for his comments on the literature .
99 d. j. aldous , deterministic and stochastic models for coalescence ( aggregation and coagulation ) : a review of the mean - field theory for probabilists , _ bernoulli _ * 5 * ( 1999 ) , pp . 348 . d.j .
aldous and j. pitman , the standard additive coalescent , , * 26 * ( 1998 ) , pp .
r. arratia , a. d. barbour and s. tavar , _ logarithmic combinatorial structures : a probabilistic approach _ , book , preprint ( 2000 ) .
e. bolthausen and a .- s .
sznitman , on ruelle s probability cascades and an abstract cavity method , , * 197 * ( 1998 ) , pp .
r. brooks , _ private communication _ ( 1999 ) .
evans and j. pitman , , , * 34 * ( 1998 ) , pp . 339383 .
a. gnedin and s. kerov , a characterization of gem distributions via a split - and - merge transformation , _ preprint _ ( 2000 ) .
see http://mouton.stochastik.math.uni-goettingen.de/~gnedin/gem.ps i. jeon ,
existence of gelling solutions for coagulation - fragmentation equations , _ comm .
phys . _ * 194 * ( 1998 ) , pp .
. j. f. c. kingman , random discrete distributions , _
statist . soc .
b _ * 37 * ( 1975 ) , pp .
122 . j. f. c. kingman , _ poisson processes _ , oxford university press , oxford ( 1993 ) . s. p. meyn and r. l. tweedie , _ markov chains and stochastic stability _ , springer , new - york ( 1994 )
. j. pitman , random discrete distributions invariant under size
biased permutation , _ adv .
_ * 28 * ( 1996 ) , pp .
j. pitman , poisson dirichlet and gem invariant distributions for split - and - merge transformations of an interval partition , _ technical report no .
597 , department of statistics , berkeley _
( may 2001 ) .
j. pitman and m. yor , , , * 25 * ( 1997 ) , pp .
. n. v. tsilevich , stationary random partitions of positive integers , _ theor .
* 44 * ( 2000 ) , pp . | we consider markov chains on the space of ( countable ) partitions of the interval @xmath0 $ ] , obtained first by size biased sampling twice ( allowing repetitions ) and then merging the parts with probability @xmath1 ( if the sampled parts are distinct ) or splitting the part with probability @xmath2 according to a law @xmath3 ( if the same part was sampled twice ) .
we characterize invariant probability measures for such chains .
in particular , if @xmath3 is the uniform measure then the poisson - dirichlet law is an invariant probability measure , and it is unique within a suitably defined class of `` analytic '' invariant measures .
we also derive transience and recurrence criteria for these chains . |
# 1 -0.8 cm [ [ section-1 ] ] # 1 ~~ # 1figure [ # 1 ] # 1figure [ # 1 ] # 1table [ # 1 ] # 1table [ # 1 ] # 1#1 * c * = cmmib10 *l * *p * * u *
the release of the first results @xcite from the wilkinson microwave anisotropy probe ( wmap ) constituted a major milestone in cosmology , laying a solid foundation upon which to found the cosmological quest in coming years .
although much of the attention in the wake of the wmap release has focused on `` sexy '' issues like the power spectrum and its cosmological implications , the primary stated science goal of wmap is to produce maps . indeed ,
one of the qualitatively new types of research made possible by wmap involves taking advantage of this spatial information by cross - correlating the maps with other cosmological templates such as galaxy @xcite , x - ray @xcite , infrared and lensing maps @xcite , which can reveal interesting signals ranging from the late integrated sachs - wolfe effect to the sz effect and lensing @xcite .
many such future studies will be looking for signals of modest statistical significance , so it is important to quantify and minimize unwanted signals in the map due to foreground contamination and detector noise . accurately understanding the foreground signal is also important for the interpretation of the wmap early reionization detection @xcite and for the interpretation of the low wmap quadrupole @xcite , since galactic foregrounds are most important on large angular scales @xcite .
the wmap team has already performed a careful foreground analysis @xcite combining the five frequency bands into a single foreground - cleaned map , shown in figure 1 ( top ) . given the huge effort that has gone into creating the spectacular multifrequency maps
, it is clearly worthwhile to subject them to an independent foreground analysis .
this is the purpose of the present paper , which we will argue not only corroborates the findings of the wmap team , but also makes some further improvements that we believe are useful .
the main goal of this paper is to remove foregrounds , not to understand or model them . for reviews of foreground modeling issues ,
see , , @xcite and references therein .
there is a rich literature of techniques for foreground removal .
the work most closely related to the present paper is that done in preparation for the planck mission @xcite , developing multipole - based cleaning techniques and testing them on simulations .
= 15.1 cm -26.05 cm = 14.7 cm -7.42 cm = 14.7 cm
-7.42 cm = 14.7 cm -3 cm = 17.5 cm
a key goal of the cmb community is to measure the function @xmath4 , the true cmb sky temperature in the sky direction given by the unit vector @xmath5 .
the wmap team have observed the sky in five frequency bands denoted k , ka , q , v and w , centered on the frequencies of 22.8 , 33.0 , 40.7 , 60.8 and 93.5 ghz , respectively , producing five corresponding maps ( figure 2 , left ) that we will refer to as @xmath6 , @xmath7 . in practice , each of these maps are discretized into @xmath8,@xmath9,@xmath10 healpix pixels @xcite , so @xmath6 is an @xmath11-dimensional vector .
however , since the maps are more than adequately oversampled relative to the beam resolution , it is equivalent and often simpler to think of them simply as five smooth functions @xmath12 .
conversely , we will occasionally write the true sky as a pixelized vector @xmath13 .
these maps are related to the true sky @xmath13 by the affine relation _
i = _ i+_i , where the matrix @xmath14 encodes the effect of beam smoothing and the additive term @xmath15 is the contribution from detector noise and foreground contamination , collectively referred to as `` junk '' below since it complicates the recovery of @xmath13 . defining the larger matrices and vectors ( let us close by returning from small to large angular scales .
the surprisingly small cmb quadrupole has intrigued the cosmology community ever since if was first observed by cobe / dmr @xcite , and simulations by @xcite show that the low value observed by wmap is sufficiently unlikely to warrant serious concern .
the wmap team measured the quadrupole using only the part of the sky outside of their galactic cut , and stress that the dominant uncertainty in its value is foreground modeling .
while we agree with this assessment , it should be borne in mind that noise variance and beam issues are completely negligible on these huge angular scales , so this does not imply that the foreground uncertainties are large compared to the signal itself . indeed
, suggests that foregrounds are subdominant to the intrinsic cmb signal even _ without any galaxy cut _ if a foreground - cleaned map is used . while the reader may feel disturbed by the clearly visible galactic residuals in figure 1 ( in both the top and bottom maps ) ,
it is important to bear in mind that these signals are present in only a tiny fraction of the total sky area and therefore contribute little to the total power spectrum . to quantify the impact of the galactic plane , we computed a sequence of all - sky power spectra for both our map and the ilc map .
we first used the unadulterated maps , then repeated the calculation after repacing the dirtiest of the nine regions from with zeroes , then zeroed our the 2nd dirtiest region as well , this zeroing procedure obviously biases the measurements by removing power , by an amount related to the zeroed area , but provides a powerful test of how sensitive the results are to galactic plane details .
we also band - pass filtered the resulting maps to produce spatial plots of the quadrupole , octopole and hexadacapole .
we found that zeroing out the dirtiest parts of the galactic plane had a negligible effect on both the power spectrum and on the spatial structure of these lowest multipoles .
specifically , we could zero out all but the three cleanest regions ( everything with @xmath16mk ) without the quadrupole or octopole changing substantially .
the spatial morphology of the quadrupole , octopole and hexadecapole for the all - sky analysis agrees well between our map and the ilc map , again showing insensitivity to galaxy modeling details ( in particular , the ilc map is likely to have less contamination in the galactic plane due to more subdivisions there ) .
although more detailed foreground modeling would be needed to rigorously quantify the foreground contribution to low multipoles , let us , encouraged by the above - mentioned tests , tentatively assume that this contribution is unimportant and perform an all - sky analysis of our cleaned map .
the resulting power spectrum is shown in , which is simply a blow - up of the leftmost part of the previous figure , and the corresponding angular correlation function is shown in .
table 1 summarizes the quadrupole and octopole results .
we see that although the quadrupole is still low , it is not quite as low as that from the cut - sky wmap team analysis of @xcite .
moreover , our map has a quadrupole virtually identical to the wmap team ilc map despite the differences in foreground modeling , further supporting our hypothesis that the quadrupole is not strongly affected by foregrounds .
the second column in table 1 shows the probability of the quadrupole in our hubble volume being as low as observed if the best - fit wmap team model from @xcite is correct .
this is computed for the all - sky case where @xmath17 has a @xmath18-distribution with 5 degrees of freedom , so the probability tabulated is simply @xmath19/\gamma(5/2)$ ] , where @xmath20 and @xmath21 are the incomplete and complete gamma functions , respectively using markov chains . ] .
we see that the statistical significance of the low quadrupole problem drops substantially with our all - sky analysis , below the 95% significance level , requiring merely a one - in - twenty fluke . to understand why the all - sky quadrupole is larger than the cut - sky one measured by the wmap team
, we plot the lowest three multipoles of our cleaned map in , all on the same temperature scale .
several features are noteworthy .
first of all , the quadrupole is low with an interesting alignment .
a generic quadrupole has three orthogonal pairs of extrema ( two maxima , two minima and two saddle points ) .
we see that the actual cmb quadrupole has its strongest pair of lobes , apparently coincidentally , fall near the galactic plane .
applying a galaxy cut therefore removes a substantial fraction of the quadrupole power .
the saddle point is seen to be close to zero . in other words
, there is a preferred axis in space along which the observed quadrupole has almost no power .
second , the observed quadrupole is the sum of the cosmic quadrupole and the dynamic quadrupole due to our motion relative to the cmb rest frame - see @xcite for a detailed discussion .
since this motion is accurately known from the cmb dipole measurement @xcite , the dynamic quadrupole can and should be subtracted when studying the cosmic contribution . shows the dynamic quadrupole approximated by @xmath22 , where @xmath23 towards @xmath24 is the velocity of the solar system relative to the cmb @xcite and @xmath25 is the angle relative to this velocity vector .
this is a small correction with peak - to - peak amplitude @xmath26 , and table 1 shows that it reduces the cosmic quadrupole slightly .
this approximation is crude since the dynamic quadrupole in fact depends on frequency @xcite , and therefore may have been either over- or under - estimated in our foreground cleaned map - we have shown it here merely to illustrate its spatial orientation and give an order - of - magnitude estimate of its importance .
third , although the overall octopole power is large , not suppressed like the quadrupole , it too displays the unusual property of a preferred axis along which power is suppressed .
moreover , this axis is seen to be approximately aligned with that for the quadrupole .
the reason that our measured octopole in is larger than that reported by the wmap team is therefore , once again , that much of the power falls within the galaxy cut .
in contrast , the hexadecapole is seen to exhibit the more generic behavior we expect of an isotropic random field , with no obvious preferred axis .
how significant is this quadrupole - octopole alignment ? as a simple definition of preferred axis for an arbitrary sky map , @xcite computes the unit vector @xmath5 around which the angular momentum dispersion @xmath27 is maximized . here
@xmath28 denotes the spherical harmonic coefficients of the map in a rotated coordinate system with its @xmath29-axis in the the @xmath5-direction .
the preferred axes @xmath30 and @xmath31 for the quadrupole and octopole , respectively , are _ 2&= & ( -0.1145 , -0.5265 , 0.8424 ) , + _ 3&= & ( -0.2578 , -0.4207 , 0.8698 ) , , both roughly in the direction of @xmath1 in virgo .
the angle between these two axes is merely @xmath32 , and the probability that a random axis falls inside a circle of radius @xmath32 around the quadrupole axis is simply the area of this circle over the area of the half - sphere , about @xmath33 . in other words , if the cmb is an isotropic gaussian random field , then a chance alignment this good requires about a 1-in-62 fluke .
this issue is discussed in greater detail in @xcite .
what does this all mean ?
although we have presented these low multipole results merely in an exploratory spirit , and more thorough modeling of the foreground contribution to @xmath34 and @xmath35 is certainly warranted , it is difficult not to be intrigued by the similarities of with what is expected in some non - standard models , for instance ones involving a flat `` small universe '' with a compact topology as in @xcite and one of the three dimensions being relatively small ( of order the horizon size or smaller )
. this could have the effect of suppressing the large - scale power in this particular spatial direction in the same sense as is seen in .
as so often in science when measurements are improved , wmap has answered old questions and raised new ones .
we thank the the wmap team for producing such a superb data set and for promptly making it public via the legacy archive for microwave background data analysis ( lambda ) .
support for lambda is provided by the nasa office of space science .
we thank krzysztof grski and collaborators for creating the healpix package @xcite , which we used both for harmonic transforms and map plotting .
thanks to ed bertschinger , gary hinshaw , jim peebles , ned wright and matias zaldarriaga for helpful comments .
this work was supported by nsf grants ast-0071213 , ast-0134999 , & ast-0205981 , and by nasa grants nag5 - 10763 & nag5 - 11099 .
mt was supported by a david and lucile packard foundation fellowship and a cottrell scholarship from research corporation .
bennett c l ; 2003;astro - ph/0302207 bennett c l ; 2003;astro - ph/0301158 bennett c l ; 2003;astro - ph/0302208 barnes c ; 2003;astro - ph/0301159 barnes c ; 2003;astro - ph/0302215 hinshaw g ; 2003;astro - ph/0302217 hinshaw g ; 2003;astro - ph/0302222 jarosik n ; 2003;astro - ph/0301164 jarosik n ; 2003;astro - ph/0302224 kogut a ; 2003;astro - ph/0302213 komatsu e ; 2003;astro - ph/0302223 limon m ; 2003;astro - ph/ page l ; 2003;astro - ph/0301160 page l ; 2003;astro - ph/0302214 page l ; 2003;astro - ph/0302220 peiris h v ; 2003;astro - ph/0302225 spergel d n ; 2003;astro - ph/0302209 verde l ; 2003;astro - ph/0302218 peiris h vspergel d n;2000;apj;540;605 boughn s p crittenden r gkoehrsen g p;2002;apj;580;672 diego j msilk j;2003;astro - ph/0302268 song y s cooray a knox lzaldarriaga m;2002;apj;590;664 cooray a;2002;prd;65;103510 hirata c mseljak u;2003;prd;67;043001 haiman zholder g p;2003;astro - ph/0302403 holder g haiman z kaplinghat mknox l;2003;astro - ph/0302404 hui lhaiman z;2003;astro - ph/0302439 tegmark mefstathiou g;1996;mnras;281;1297 de oliveira - costa a ; 2002;apj;567;363 bouchet f r ; 1995;space science rev.;74;37 bouchet f rgispert r;1999;new astron.;4;443 bouchet f r , gispert rpuget j l;1996 ; unveiling the cosmic infrared background;e .
dwek;aip;baltimore;255 bouchet f r , prunet ssethi s k;1999;mnras;302;663 tucci m , carretti e , cecchini s , fabbri r , orsini mpierpaoli e;2000;new astr;5;181 baccigalupi c , burigana c , perrotta f , de zotti g , la porta l , maino d , maris mpaladini r;2001;a&a;372;8 burigana cla porta l;2002;astro - ph/0202439 bruscoli m , tucci m , natale v , carretti e , fabbri r , sbarra ccortiglioni s;2002;new astr;7;171
grski k m , hivon ewandelt b d;1999 ; astro - ph/9812350;evolution of large - scale structure;banday a j , sheth r s l;printpartners ipskamp;nethernalds;37 grski k m , wandelt b d , hansen f kbanday a j;1999;astro - ph/9905275 wright e l ; 1996;astro - ph/9601059 tegmark m;1997;apjl;480;l87 muciaccia p f , natoli pvittorio n;1997;apj;488;63 knox l;1995;phys .
d;48;3502 tegmark m;1998;apj;502;1 hivon e , grski k m , netterfield c b , crill b p , prunet shansen f;2002;apj;567;2 ruhl j e;2002;astro - ph/0212229 tegmark m;1997;phys .
d;55;5895 wiener n;1949;extrapolation and smoothing of stationary time series;new
york;wiley rybicki g bpress w h;1992;apj;398;169 lahav o ; 1994;apj;423;l93 fisher k b ; 1995;mnras;272;885 zaroubi s ; 1995;apj;449;446 bond j r;1995;phys .
lett.;74;4369 bunn e f , hoffmann ysilk j;1996;apj;464;1 tegmark m ; 1996;apjl;474;l77 mather j ; 1994;apj;420;439 starobinsky n n;1993;jetp lett.;57;622 stevens d , scott dsilk j;1993;prl;71;20 asmoot g f;1995;apj;448;477 a , smoot g fstarobinski a a;1996;apj;468;457 levin j;2002;phys .
rep.;365;251 rocha g ; 2002;astro - ph/0205155 | we perform an independent foreground analysis of the wmap maps to produce a cleaned cmb map ( available online ) useful for cross - correlation with , , galaxy and x - ray maps .
we use a variant of the tegmark & efstathiou ( 1996 ) technique that assumes that the cmb has a blackbody spectrum , but is otherwise completely blind , making no assumptions about the cmb power spectrum , the foregrounds , wmap detector noise or external templates . compared with the foreground - cleaned internal linear combination map produced by the wmap team , our map has the advantage of containing less non - cmb power ( from foregrounds and detector noise ) outside the galactic plane .
the difference is most important on the the angular scale of the first acoustic peak and below , since our cleaned map is at the highest ( 12.6@xmath0 ) rather than lowest ( 49.2@xmath0 ) wmap resolution .
we also produce a wiener filtered cmb map , representing our best guess as to what the cmb sky actually looks like , as well as cmb - free maps at the five wmap frequencies useful for foreground studies .
we argue that our cmb map is clean enough that the lowest multipoles can be measured without any galaxy cut , and obtain a quadrupole value that is slightly less low than that from the cut - sky wmap team analysis .
this can be understood from a map of the cmb quadrupole , which shows much of its power falling within the galaxy cut region , seemingly coincidentally .
intriguingly , both the quadrupole and the octopole are seen to have power suppressed along a particular spatial axis , which lines up between the two , roughly towards @xmath1 in virgo .
# 1 # 1 # 2#2 .
# 1 # 1 # 2 # 3#2 .
# 3 .
# 1 # 1 # 2 # 3 # 4#2 .
# 3 .
# 4 # 1 # 1 # 2 # 3 # 4 # 5#2 .
# 3 .
# 4 # 5 .
# 1 # 1;#2;#3;#4;#5 # 1 , # 3 * # 4 * , # 5 ( # 2 ) . # 1;#2;#3;#4;#5;#6 # 1 , # 3 * # 4 * , # 5 ( # 2 ) . # 1;#2;#3;#4;#5 # 1 , _ # 3 _ ( # 5 , # 4 , # 2 ) .
# 1;#2;#3 # 1 , # 3 ( # 2 ) .
# 1;#2;#3;#4;#5;#6 # 1 # 2 , in _ # 3 _ , ed .
# 4 ( # 5 : # 6 ) # 1;#2;#3;#4;#5;#6;#7 # 1 # 2 , in _ # 3 _ , ed . # 4 ( # 5 : # 6 ) , p#7 # 1@xmath2 # 1@xmath3 # 1equation ( [ # 1 ] ) # 1equation ( [ # 1 ] ) # 1 ( [ # 1 ] ) # 1to 0pt#1 # 1@xmath2 # 1section [ # 1 ] # 1section [ # 1 ] # 11 2.3truecm 11.45truecm#1 # 1figure [ # 1 ] # 1 -1.0 cm |
recent solar @xcite , atmospheric@xcite , reactor@xcite and accelerator@xcite neutrino oscillation experiments have provided us with very robust evidence that neutrinos are massive and lepton flavors are mixed . the canonical seesaw mechanism @xcite gives a very simple and appealing explanation of the smallness of left - handed neutrino masses it is attributed to the largeness of right - handed neutrino masses .
the existence of neutrino oscillations implies the violation of lepton flavors .
hence the lepton flavor violating ( lfv ) decays in the charged - lepton sector , such as @xmath3 , should also take place .
they are unobservable in the standard model ( sm ) , because their decay amplitudes are expected to be highly suppressed by the ratios of neutrino masses ( @xmath4 ev ) to the @xmath5-boson mass ( @xmath6 gev ) . in the supersymmetric extension of the sm , however , the branching ratios of such rare processes can be enormously enlarged .
current experimental bounds on the lfv decays @xmath7 , @xmath8 and @xmath9 are@xcite @xmath10 the sensitivities of a few planned experiments@xcite may reach @xmath11 , @xmath12 and @xmath13 . for simplicity ,
here we work in the framework of the minimal supergravity ( msugra ) extended with two heavy right - handed majorana neutrinos . then all the soft breaking terms are diagonal at high energy scales , and the only source of lepton flavor violation in the charged - lepton sector is the radiative correction to the soft terms through the neutrino yukawa couplings . in other words ,
the low - energy lfv processes @xmath14 are induced by the rge effects of the slepton mixing .
the branching ratios of @xmath14 are given by @xcite @xmath15 ^ 2 |c^{}_{ij}|^2 \tan^2 \beta \ ; , % ( 2)\ ] ] where @xmath16 and @xmath17 denote the universal scalar soft mass and the trilinear term at @xmath18 , respectively .
in addition , @xmath19 is a typical mass of superparticles , can be approximately written as @xcite @xmath20 with @xmath21 being the gaugino mass ; and @xmath22 with @xmath23 gev to be fixed in our calculations .
@xmath24 and @xmath25 ( for @xmath26 ) represent the dirac neutrino mass matrix and the masses of the heavy right - handed majorana neutrinos , respectively .
to calculate the branching ratios of @xmath14 , we need to know the following parameters in the framework of the msugra : @xmath21 , @xmath16 , @xmath17 , @xmath27 and @xmath28 .
these parameters can be constrained from cosmology ( by demanding that the proper supersymmetric particles should give rise to an acceptable dark matter density ) and low - energy measurements ( such as the process @xmath29 and the anomalous magnetic moment of muon @xmath30 ) .
here we adopt the snowmass points and slopes ( sps ) @xcite listed in table .
1 . these points and slopes are a set of benchmark points and parameter lines in the msugra parameter space corresponding to different scenarios in the search for supersymmetry at present and future experiments .
since the canonical seesaw models are usually pestered with too many parameters , specific assumptions have to be made for @xmath24 or @xmath25 so as to calculate the lfv rare decays @xcite .
in addition , the lfv processes have also been discussed in the supersymmetric minimal seesaw model ( msm ) @xcite , in which only two heavy right - handed majorana neutrinos are introduced @xcite . in the msm ,
all model parameters can in principle be fixed by use of the lfv rare decays and eletric dipole moment of the electron @xcite .
ibarra and ross discuss that the lfv processes may constrain the masses of right - handed heavy majorana neutrinos @xcite . in the frampton - glashow - yanagida @xcite ( fgy )
ansatz , raidal and strumia have given analytic approximations to the lfv rare decays for the normal hierarchy case @xcite . as shown in the literature @xcite , the fgy ansatz only includes three unknown parameters : @xmath31 , @xmath32 and the smallest mixing angle @xmath33 . in this paper
, we are going to numerically compute the lfv processes in terms of @xmath33 for both normal and inverted hierarchies in the fgy ansatz . in the following parts
, we shall show that the branching ratio of @xmath0 may distinguish the neutrino mass hierarchies , the jarlskog parameter of cp violation ( @xmath34 ) and the effective mass of neutrinoless double beta decay ( @xmath35)@xcite . ] .
in addition , once the ratio of @xmath1 to @xmath2 is measured , one may distinguish the normal and inverted neutrino mass hierarchies , and confirm the masses of two heavy right - handed majorana neutrinos by means of the lfv processes and the thermal leptogenesis mechanism@xcite .
it is worthwhile to stress that this conclusion is independent of the msugra parameters .
the remaining parts of this paper are organized as follows . in section
ii , we briefly describe the main features of the fgy ansatz and the thermal leptogenesis in the minimal supersymmetric standard model ( mssm ) . in section iii ,
the branching ratio of @xmath0 is computed in terms of the sps . in section iv
, we numerically calculated the ratio of @xmath36 to @xmath37 .
finally the summary are given in section v.
in the supersymmetric extension of the msm , two heavy right - handed majorana neutrinos @xmath38 ( for @xmath39 ) are introduced as the @xmath40 singlets . the lagrangian relevant for lepton masses
can be written as @xcite @xmath41 where @xmath42 denotes the left - handed lepton doublet , while @xmath43 and @xmath44 stand respectively for the right - handed charged - lepton and neutrino singlets . @xmath45 and @xmath46 ( with hypercharges @xmath47 ) are the mssm higgs doublet superfields . after the spontaneous gauge symmetry breaking , one obtains the charged - lepton mass matrix @xmath48 and the dirac neutrino mass matrix @xmath49 . here
@xmath50 being the vev of the higgs doublet @xmath51 ( for @xmath52 ) .
an important parameter @xmath53 is defined by @xmath54 or @xmath55 with @xmath56 .
the heavy right - handed majorana neutrino mass matrix @xmath57 is a @xmath58 symmetric matrix and @xmath59 is a @xmath60 matrix . without loss of generality
, we work in the flavor basis where @xmath61 and @xmath57 are both diagonal , real and positive ; i.e. , @xmath62 and @xmath63 .
the seesaw relation @xcite @xmath64 remains valid .
note that this canonical seesaw relation holds up to the accuracy of @xmath65 @xcite . since @xmath57 is of rank 2
, @xmath66 is also a rank-2 matrix with @xmath67 , where @xmath68 ( for @xmath69 ) are the masses of three light neutrinos . as for three neutrino masses
@xmath70 , the solar neutrino oscillation data have set @xmath71 @xcite .
now that the lightest neutrino in the msm must be massless , we are then left with either @xmath72 ( normal mass hierarchy ) or @xmath73 ( inverted mass hierarchy ) . with the best - fit values @xmath74 and @xmath75 @xcite , one can numerically calculate the neutrino masses @xcite . in the fgy ansatz @xcite
, @xmath24 is taken to be of the form @xmath76 with the help of eqs.(6 ) and ( 7 ) , one may straightforwardly arrive at @xmath77 without loss of generality , one can always redefine the phases of left - handed lepton fields to make @xmath78 , @xmath79 and @xmath80 real and positive .
in this basis , only @xmath81 is complex and its phase @xmath82 is the sole source of cp violation in the model under consideration .
because @xmath78 , @xmath79 and @xmath80 of @xmath24 have been taken to be real and positive , @xmath66 may be diagonalized in a more general way @xmath83 where @xmath84 and @xmath85 is lepton flavor mixing matrix @xcite parameterized as @xmath86 with @xmath87 , @xmath88 and so on .
@xmath89 , @xmath90 and @xmath91 hold as a good approximation @xcite . in view of the current experimental data , we have @xmath92 and @xmath93 ( best - fit values ) as well as @xmath94 at the @xmath95 confidence level @xcite .
it is worth remarking that there is only a single nontrivial majorana cp - violating phase ( @xmath96 ) in the msm , as a straightforward consequence of @xmath97 or @xmath98 .
note that all six phase parameters ( @xmath99 , @xmath96 , @xmath100 , @xmath101 , @xmath53 and @xmath102 ) have been determined in terms of @xmath103 , @xmath104 , @xmath105 and @xmath33 @xcite : @xmath106 \ ; , \nonumber \\ \sigma & = & \frac{1}{2 } \arctan \left [ \frac{c^{}_x s^{}_y \sin\delta } { s^{}_x c^{}_y s^{}_z + c^{}_x s^{}_y \cos\delta } \right ] \ ; , \nonumber \\ \alpha & = & -\frac{1}{2 } \arctan \left [ \frac{r^{2}_{23 } s^2_x
c^2_z \sin 2\sigma}{s^2_z + r^2_{23 } s^2_x c^2_z \cos 2\sigma } \right ] \ ; , \nonumber \\ \beta & = & -\gamma - \arctan \left [ \frac{c^{}_x c^{}_y s^{}_z \sin\delta } { s^{}_x s^{}_y - c^{}_x c^{}_y s^{}_z \cos\delta } \right ] \ ; , \nonumber \\ \gamma & = & \frac{1}{2 } \arctan \left [ \frac{s^2_z \sin 2\sigma}{r^2_{23 } s^2_x c^2_z + s^2_z \cos 2\sigma } \right ] \ ; , \nonumber \\ \phi & = & \alpha + \beta - \arctan \left [ \frac{s^{}_x c^{}_y s^{}_z \sin\delta}{c^{}_x s^{}_y + s^{}_x c^{}_y s^{}_z \cos\delta } \right ] % ( 11)\end{aligned}\ ] ] for the normal mass hierarchy ( @xmath107 ) .
similar results can also be obtained for the inverted mass hierarchy ( @xmath98 ) @xcite , but we do not elaborate on them here . because @xmath108 must hold , we find @xmath109 ( @xmath107 ) and @xmath110 ( @xmath98 ) .
hence , a measurement of the unknown angle @xmath33 becomes crucial to test the model . in the flavor basis where the mass matrices of charged leptons and right - handed neutrinos are diagonal , one can calculate the cp asymmetry in the decays of the lighter right - handed neutrino and obtain@xcite @xmath111 when @xmath112 , where we have used @xmath113 then @xmath114 can result in a net lepton number asymmetry @xmath115 .
the lepton number asymmetry @xmath115 is eventually converted into a net baryon number asymmetry @xmath116 via the nonperturbative sphaleron processes@xcite , which is given by @xmath117 where @xmath118 , @xmath119 is an effective number characterizing the relativistic degrees of freedom which contribute to the entropy @xmath120 of the early universe , and @xmath121 accounts for the dilution effects induced by the lepton - number - violating wash - out processes .
the dilution factor @xmath121 can be figured out by solving the full boltzmann equations@xcite , however , we take the following approximate formula@xcite : @xmath122^{-0.6 } % ( 15)\ ] ] with @xmath123 .
it is clear that @xmath114 and @xmath124 only involve two unknown parameters : @xmath31 and @xmath125 .
is inversely proportional to the msugra parameter @xmath126 . because @xmath127 is disfavored ( as indicated by the higgs exclusion bounds@xcite ) ,
here we focus on @xmath128 or equivalently @xmath129 .
hence @xmath130 is a reliable approximation in our discussion . ]
a generous range @xmath131 has been drawn from the recent wilkinson microwave anisotropy probe ( wmap ) observational data @xcite .
it has been shown that @xmath132 gev ( @xmath133 ) and @xmath134 gev ( @xmath135 ) are required by the current observational data of @xmath124 @xcite .
flavor effects in the mechanism of thermal leptogenesis have recently attracted a lot of attention@xcite . because the @xmath136 yukawa coupling are in thermal equilibrium for @xmath137
gev , the flavor issue should be taken into account in our model for the @xmath133 case . using the analytic approximate formulas in the ref.@xcite
, one may calculate the cp - violating asymmetry @xmath138 and the corresponding wash - out parameter @xmath139 for the @xmath136 lepton doublet in the final states of @xmath140 decays .
we find @xmath141 and @xmath142 because of @xmath143 , namely the @xmath144 decays involving the @xmath136 lepton doublet do nt contribute to the baryon asymmetry @xmath145 . on the other hand ,
the cp - violating asymmetry @xmath146 , which is produced by the out - of - equilibrium decays of @xmath147 , vanishes .
therefore , such flavor effects may be negligible in our paper .
note that there is in general a potential conflict between achieving successful thermal leptogenesis and avoiding overproduction of gravitinos in the msm with supersymmetry @xcite unless the gravitinos are heavier than @xmath148 tev @xcite .
if the mass scale of gravitinos is of @xmath149 tev , one must have @xmath150 gev .
this limit is completely disfavored in the fgy ansatz .
such a problem could be circumvented in other supersymmetric breaking mediation scenarios ( e.g. , gauge mediation @xcite ) or in a class of supersymmetric axion models @xcite , where the gravitino mass can be much lighter in spite of the very high reheating temperature . for simplicity ,
we choose @xmath151 as an input parameter in this paper .
hence , one may analyze the dependence of @xmath152 on @xmath125 from the successful leptogenesis .
with the help of eqs .
( 4 ) and ( 7 ) , @xmath153 can explicitly be written as @xmath154 because of @xmath155 , we are left with @xmath156 . if @xmath157 is established from the future experiments , it will be possible to exclude the fgy ansatz .
( 13 ) , we reexpress eq . ( 16 ) as @xmath158 as shown in section ii , @xmath31 may in principle be confirmed by leptogenesis for given values of @xmath159 , but @xmath32 is entirely unrestricted from the successful leptogenesis with @xmath160 .
we numerically calculate br(@xmath0 ) for different values of @xmath159 by using the sps points . from the qed correction @xcite . ]
the results are shown in fig .
1 . since the sps points 1a and 1b ( or points 2 and 3 ) almost have the same consequence in our scenario , we only focus on point 1a ( or point 3 ) . when @xmath161 or @xmath162 , the future experiment is likely to probe the branching ratio of @xmath163 in the @xmath164 case .
the reason is that @xmath165 ( or @xmath166 ) implies @xmath167 ( or @xmath168 ) .
furthermore , the successful leptogenesis requires a very large @xmath31 due to @xmath169 .
it is clear that all sps points are all unable to satisfy @xmath170 in the @xmath171 case .
therefore , we can exclude the @xmath98 case when the sps points are taken as the msugra parameters .
when @xmath172 , @xmath173 arrives at its minimal value in the @xmath171 case . for the sps slopes , larger @xmath21 yields smaller @xmath37
we plot the numerical dependence of br(@xmath0 ) on @xmath21 in fig .
2 , where we have adopted the sps slope 3 and taken @xmath174 .
we find that @xmath175 ( or @xmath176 ) can result in @xmath177 for @xmath178 ( or @xmath179 ) .
for all values of @xmath21 between 300 gev and 1000 gev , @xmath37 is larger than the sensitivity of some planned experiments , which ought to examine the @xmath98 case when the sps slope 3 is adopted . the same conclusion can be drawn for the sps slopes 1a and 2 . in view of the present experimental results on muon @xmath180
, one may get @xmath181 gev for @xmath182 and @xmath183@xcite , implying that the @xmath98 case should be disfavored .
in this paper , an important conclusion is the masses of two heavy right - handed majorana neutrinos ( @xmath152 and @xmath184 ) can be derived through the leptogenesis and the lfv rare decays . with the help of eqs.(2 ) and ( 17 ) , one can obtain @xmath185 ^ 2 } { m_1 ^ 2 \ ; \ ; .
% ( 18)\end{aligned}\ ] ] since the successful leptogenesis can be used to fix @xmath31 , a measurement of the above ratio will allow us to determine or constrain @xmath32 . on the other hand , we may calculate the ratio @xmath186 by inputting the appropriate @xmath184 .
it is worthwhile to remark that the ratio @xmath186 is independent of the msugra parameters .
to illustration , we show the numerical results of @xmath186 as functions of @xmath187 in fig .
3(a ) for the @xmath188 case and fig .
3(b ) for the @xmath189 case , respectively . in the @xmath135 case ,
when @xmath190 , the phase @xmath168 which implies @xmath152 and @xmath184 may be larger than the gut scale @xmath191 gev .
so we demand that @xmath152 and @xmath184 must be less than the gut scale . below the scale @xmath18 ,
@xmath192 obtain its maximum value at @xmath193 consequently , we can derive the maximum value of @xmath186 with @xmath194 gev . in fig .
3(b ) , the numerical results at the left side of the red ( green ) dash - dot line are nonphysical " , because these results are corresponding to @xmath195 gev for the @xmath196 ( @xmath197 ) case . we can obtain the ratio @xmath198 once @xmath186 is measured , furthermore calculate the heavy neutrino mass @xmath184 . from fig .
3 , one may derive @xmath199 for the @xmath200 case and @xmath201 for the @xmath202 case , respectively .
if the future experiments prove @xmath203 , the inverted mass hierarchy case can be excluded .
it is worthwhile to stress that the above conclusions are independent of the msugra parameters .
finally , let us comment on the @xmath204 case @xcite . in this case , the cp - violating asymmetry @xmath205 , which is produced by the out - of - equilibrium decay of @xmath147 , may finally survive . in order to producing the positive cosmological baryon number asymmetry ( @xmath206 ) , @xmath99 of eq.(11 )
must take - sign .
the flavor effects may be neglected since the successful thermal leptogenesis requires @xmath207 gev ( @xmath107 ) and @xmath208 gev ( @xmath209 ) . for @xmath210
, we calculate the branching ratio of @xmath0 , find the inverted hierarchy case is disfavored for all sps points and slopes . for the normal hierarchy case
, on may distinguish the @xmath211 case ( @xmath212 ) and the @xmath204 case ( @xmath213 ) in terms of the ratio @xmath186 .
in addition , the masses of heavy neutrinos may be quasi - degenerate @xcite . in the @xmath214 case
, the so - called resonant leptogenesis may occur @xcite .
such a scenario could allow us to relax the lower bound on the lighter right - handed majorana neutrino mass .
therefore , the branching ratios of @xmath0 and @xmath215 may be much less than the sensitivities of a few planned experiments .
@xmath216 ( normal hierarchy ) and @xmath217 ( inverted hierarchy ) .
we have analyzed the lfv rare decays in the supersymmetric version of the msm in which the fgy ansatz is incorporated . in this scenario , there are only three unknown parameters @xmath125 , @xmath152 and @xmath184 .
the successful leptogenesis may fix @xmath152 for given values of @xmath125 .
we have numerically calculated the branching ratio of @xmath0 for different values of @xmath218 by using the sps points .
then one can find the inverted mass hierarchy case is disfavored by all sps points .
for the sps slopes , the planned experiments may measure @xmath219 in the @xmath135 case .
in addition , once the ratio of @xmath220 to @xmath221 is measured , we may also distinguish the neutrino mass hierarchies and fix the masses of two heavy right - handed majorana neutrinos by means of the lfv processes and the leptogenesis mechanism .
it is worthwhile to stress that this conclusion is independent of the msugra parameters .
p. minkowski , phys .
b * 67 * , 421 ( 1977 ) ; t. yanagida , in _ proceedings of the workshop on unified theory and the baryon number of the universe _ , edited by o. sawada and a. sugamoto ( kek , tsukuba , 1979 ) ; m. gell - mann , p. ramond , and r. slansky , in _ supergravity _ , edited by f. van nieuwenhuizen and d. freedman ( north holland , amsterdam , 1979 ) ; s.l . glashow , in _ quarks and leptons _ , edited by m. l@xmath222vy _ et al .
_ ( plenum , new york , 1980 ) ; r.n mohapatra and g. senjanovic , phys . rev . lett . * 44 * , 912 ( 1980 ) . mega collaboration , m.l .
et al . _ ,
83 * , 1521 ( 1999 ) ; babar collaboration , b. aubert _ et al .
lett . * 96 * , 041801 ( 2006 ) ; babar collaboration b. aubert _ et al .
lett . * 95 * , 041802 ( 2005 ) .
petcov , w. rodejohann , t. shindou and y. takanishi , nucl . phys . *
b739 * , 208 ( 2006 ) ; l. calibbi , a. faccia , a. masiero and s.k .
vempati , hep - ph/0605139 ; s.t .
petcov and t. shindou , hep - ph/0605151 ; t. ota and w. rodejohann , phys .
b * 639 * , 322 ( 2006 ) ; k.a .
hochmuth and w. rodejohann , hep - ph/0607103 ; s. antusch , e. arganda , m.j .
herrero and a. teixeira , hep - ph/0607263 ; g.c .
branco , a.j .
buras , s. jager , s. uhlig and a. weiler , hep - ph/0609067 .
a. kleppe , _ `` extending the standard model with two right - handed neutrinos '' , in * lohusalu 1995 , neutrino physics * , 118 - 125 _ ; a. kleppe , in _ proceedings of the workshop on what comes beyond the standard model _ , bled , slovenia , 29 june - 9 july 1998 ; e. ma , d.p .
roy and u. sarkar , phys .
b * 444 * , 391 ( 1998 ) ; s.f .
king , nucl .
b576 * , 85 ( 2000 ) ; jhep * 0209 * , 011 ( 2002 ) ; l. lavoura and w. grimus , jhep * 0009 * , 007 ( 2000 ) ; t. endoh , s. kaneko , s.k .
kang , t. morozumi and m. tanimoto , phys .
89 * , 231601 ( 2002 ) ; s. raby , phys .
b * 561 * , 119 ( 2003 ) ; r. kuchimanchi and r.n .
mohapatra , phys .
b * 552 * , 198 ( 2003 ) ; v. barger , d.a .
dicus , h.j .
he and t.j .
li , phys .
b * 583 * , 173 ( 2004 ) ; s. chang , s.k .
kang and k. siyeon , phys .
b * 597 * , 78 ( 2004 ) ; w. rodejohann , eur .
j. c * 32 * , 235 ( 2004 ) ; h. k. dreiner , h. murayama and m. thormeier , nucl . phys . * b729 * , 278 ( 2005 ) . m. a. luty , phys . rev .
d * 45 * , 455 ( 1992 ) ; m. flanz , e.a .
paschos and u. sarkar , phys .
b * 345 * , 248 ( 1995 ) ; l. covi , e. roulet and f. vissani , phys .
b * 384 * , 169 ( 1996 ) ; m. flanz , e.a .
paschos , u. sarkar and j. wess , phys .
b * 389 * , 693 ( 1996 ) ; m. pl@xmath224macher , z. phys .
c * 74 * , 549 ( 1997 ) ; w. buchm@xmath224ller and m. pl@xmath224macher , phys .
b * 431 * , 354 ( 1998 ) ; m. pl@xmath224macher , nucl .
b530 * , 207 ( 1998 ) ; t. hambye , nucl . phys . * b633 * , 171 ( 2002 ) ; w. buchm@xmath224ller , p.di bari and m. pl@xmath224macher , nucl . phys . *
b643 * , 367 ( 2002 ) ; g.f .
giudice , a. notari , m. raidal , a. riotto and a. strumia , nucl . phys . *
b685 * , 89 ( 2004 ) .
kolb and m.s .
turner , _ the early universe _ , addison - wesley ( 1990 ) ; h.b .
nielsen and y. takanishi , phys .
b * 507 * , 241 ( 2001 ) ; e.kh .
akhmedov , m. frigerio and a.yu .
smirnov , jhep * 0309 * , 021 ( 2003 ) ; w.l .
guo , phys .
d * 70 * , 053009 ( 2004 ) ; z.z .
xing , phys .
d * 70 * , 071302 ( 2004 ) .
r. barbieri , p. creminelli , a. strumia and n. tetradis , nucl .
b575 * , 61 ( 2000 ) ; a. abada , s. davidson , f.x .
josse - michaux , m. losada and a. riotto , jcap * 0604 * , 004 ( 2006 ) ; e. nardi , y. nir , e. roulet and j. racker , jhep * 0601 * , 164 ( 2006 ) .
.some parameters for the snowmass points and slopes ( sps ) in the msugra .
the masses are given in gev .
@xmath225 appeared in the higgs mass term has been taken as @xmath226 for all sps .
[ cols="^,^,^,^,^,<",options="header " , ] on @xmath159 : ( a ) in the @xmath97 case ; and ( b ) in the @xmath98 case . the black solid line and black dash - dot line denote the present experimental upper bound on and the future experimental sensitivity to @xmath227 , respectively.,title="fig:",width=377,height=377 ] on @xmath159 : ( a ) in the @xmath97 case ; and ( b ) in the @xmath98 case .
the black solid line and black dash - dot line denote the present experimental upper bound on and the future experimental sensitivity to @xmath227 , respectively.,title="fig:",width=377,height=377 ] on @xmath21 for sps slope 3 in the @xmath98 case .
the black solid line and black dash - dot line denote the present experimental upper bound on and the future experimental sensitivity to @xmath37 , respectively.,width=377,height=377 ] on @xmath228 : ( a ) in the @xmath97 case ; and ( b ) in the @xmath229 case.,title="fig:",width=377,height=377 ] on @xmath228 : ( a ) in the @xmath97 case ; and ( b ) in the @xmath229 case.,title="fig:",width=377,height=377 ] | we investigate the lepton flavor violating ( lfv ) rare decays in the supersymmetric minimal seesaw model in which the frampton - glashow - yanagida ansatz is incorporated .
the branching ratio of @xmath0 is calculated in terms of the snowmass points and slopes ( sps ) .
we find that the inverted mass hierarchy is disfavored by all sps points .
in addition , once the ratio of @xmath1 to @xmath2 is measured , one may distinguish the normal and inverted neutrino mass hierarchies , and confirm the masses of heavy right - handed majorana neutrinos by means of the lfv processes and the thermal leptogenesis mechanism .
it is worthwhile to stress that this conclusion is independent of the supersymmetric parameters . |
fig.1 . successive inclusion of the ope terms .
the helix lines denote the current ( [ 28 ] ) .
solid lines denote the quarks .
thick lines show the matter .
( a ) the local vector and scalar condensates are included ( contribution of the gluon condensate is not shown ) .
( b ) the factorized part of the four - quark condensates is included .
one of the @xmath432 operators acts on vacuum , while another @xmath432 operator acts inside the nucleon .
small circles stand for averaging over vacuum .
( c ) the internal parts of the four - quark operators are included .
all the quark operators act inside the nucleon .
( d ) the nonlocality of the vector condensate is included .
the dark blobs denote the nonlocal condensates .
the analogs of the ope terms in the meson - exchange interaction of the nucleon with matter .
figures 2(a - d ) correspond to the figures 1(a - d ) .
solid line denotes the nucleon .
thick lines show the matter .
wavy and dashed lines stand for vector and scalar mesons .
( a ) exchanges by vector and scalar ( `` effective '' ) mesons with the pion vertices .
( b ) exchanges by these mesons with the anomalous lorentz structures of the vertices ( the dark squares ) .
( c ) exchanges by the @xmath432 pairs with strong correlations between them , presumably the local interactions with two mesons .
double line denotes the two - meson systems .
( d ) inclusion of the nonlocal structure of the vertices of interaction between the vector mesons and the nucleons of the matter .
the dark blobs denote the formfactors of the vertices .
the density dependence of the vector self - energy @xmath0 ( solution of eqs .
( [ 57])-([59 ] ) ) .
the solid line shows the results for symmetric matter ( @xmath500 ) .
the dashed and dotted lines show the proton and neutron self - energies @xmath501 and @xmath502 in the neutron matter ( @xmath438 ) .
the density dependence of the effective mass @xmath109 ( solution of eqs .
( [ 57])-([59 ] ) ) .
the solid line shows the results for symmetric matter ( @xmath500 ) .
the dashed and dotted lines show the proton and neutron self - energies @xmath503 and @xmath504 in the neutron matter ( @xmath438 ) .
the density dependence of the functions @xmath505 , @xmath506 , @xmath507 and @xmath508 introduced by eqs .
( [ 112 ] ) , @xmath509 .
the dotted lines demonstrate the quality of fitting with the simple functions on @xmath66 , as described in appendix d. fig.6 .
the density dependence of the single particle potential energy @xmath510 , eq .
the solid line shows the results for symmetric matter ( @xmath500 ) .
the dashed and dotted lines show the proton and neutron self - energies @xmath511 and @xmath512 in the neutron matter ( @xmath438 ) . | we calculate the nucleon parameters in isospin asymmetric nuclear matter using the qcd sum rules .
the nucleon self - energies @xmath0 and @xmath1 are expressed in terms of the in - medium values of qcd condensates .
the simple approximate expressions for the self - energies are obtained in terms of these condensates .
relation between successive inclusion of the condensates and the meson - exchange picture of the nucleon interaction with medium is analyzed .
the values of the self - energies and of the symmetry energy agree with those obtained by the methods of nuclear physics .
* i. introduction * in this paper we investigate the vector and scalar self - energies of nucleons in nuclear matter composed by the neutrons and protons , distributed with densities @xmath2 and @xmath3 .
we calculate the dependence on the total density @xmath4 and on the asymmetry parameter @xmath5 ( another conventional presentation of the asymmetry parameter is @xmath6 with @xmath7 and @xmath8 standing for the total number of neutrons and protons , while @xmath9 ) .
we present also the equations for the single - particle potential energies .
the results are obtained by qcd sum rules approach .
the qcd sum rules were invented by shifman _
et al . _
@xcite to express the hadron parameters through the vacuum expectation values of qcd operators .
being initially used for the mesons , the method was expanded by ioffe @xcite to the description of the baryons . the approach succeeded in describing the static characteristics as well as some of the dynamical characteristics of the hadrons in vacuum
see , e.g. , the reviews @xcite .
the main idea is to consider the correlation function @xmath10 ( @xmath11 stands for the unit @xmath12 matrix ) describing the propagation of the system with the quantum numbers of the hadron , in the different regions of values of the momentum @xmath13 , where certain information on its behavior is available .
the asymptotic freedom of qcd enables to present @xmath14 ( @xmath15 ) at @xmath16 as the power series of @xmath17 and qcd coupling @xmath18 . on the other hand ,
the imaginary part of @xmath19 at @xmath20 can be described in terms of the observable hadrons .
this prompts to consider the dispersion relation for the function @xmath14 @xcite @xmath21 at @xmath16 .
the coefficients of the expansion of the left - hand side ( lhs ) of the functions @xmath14 in powers of @xmath17 are the expectation values of the local operators constructed of the quark and gluon fields , which are called `` condensates '' . such presentation , known as the operator product expansion ( ope ) @xcite provides the perturbative expansion of the short - distance effects , while the nonperturbative physics is contained in the condensates .
the usual treatment of the right - hand side ( rhs ) of eq .
( [ 4 ] ) consists in `` pole + continuum '' presentation , in which the lowest laying pole is singled out while the higher states are approximated by the continuum described by the pure perturbative ope contribution .
thus , eq . ( [ 4 ] ) ties the values of qcd condensates with the characteristics of the lowest hadronic state .
such interpretation requires that the contribution of the pole to the rhs of eq .
( [ 4 ] ) exceeds the contribution of the continuum .
the ope of the lhs of eq .
( [ 4 ] ) becomes increasingly true , while the value of @xmath22 increases . on the other hand ,
the `` pole + continuum '' model becomes more accurate while @xmath22 decreases .
the important assumption is that the two presentations are close in certain intermediate region of the values of @xmath13 . to improve the overlap of the qcd and the phenomenological descriptions , one usually applies certain mathematical anzats , i.e. , the borel transform .
the borel transformed dispersion relations ( [ 4 ] ) are known as qcd sum rules @xcite .
for example , the qcd sum rules for the nucleon provided a connection between the nucleon mass and the scalar quark condensate @xmath23 @xcite . a more detailed analysis requires inclusion of the expectation values of the higher dimension , e.g. , the gluon condensate @xmath24 and the four - quark condensate @xmath25 .
two other unknowns of nucleon qcd sum rules equations are the residue of the nucleon pole @xmath26 and the effective continuum threshold @xmath27 .
later the qcd sum rules were applied for investigation of modified nucleon parameters in the symmetric nuclear matter @xcite .
they were based on the borel - transformed dispersion relation for the function @xmath28 describing the propagation of the system with the quantum numbers of the nucleon ( the proton ) in the nuclear matter . considering nuclear matter as a system of @xmath29 nucleons with momenta @xmath30
, one introduces the vector @xmath31 which is thus @xmath32 in the rest frame of the matter .
the function @xmath28 can be presented as @xmath33 with the arbitrary function @xmath34 being kept constant in the dispersion relations in @xmath13 .
the general form of the function @xmath35 can thus be presented as @xmath36 the in - medium qcd sum rules are the borel - transformed dispersion relations for the components @xmath37 @xmath38 @xmath39 the spectrum of the function @xmath28 is much more complicated than that of the function @xmath40 .
the choice of the function @xmath34 is dictated by the attempts to separate the singularities connected with the nucleon in the matter from those connected with the properties of the matter itself . since the latter manifest themselves as singularities in the variable @xmath41 , the separation can be done by putting @xmath42 and by fixing @xcite @xmath43 with @xmath44 being the total nucleon energy at the fermi surface . by using eq .
( [ 8 ] ) the characteristics of the nucleon in nuclear matter can be expressed through the in - medium values of qcd condensate .
the possibility of extension of `` pole + continuum '' model @xcite to the case of finite densities was shown in @xcite@xcite .
the lowest order of ope of the lhs of eq .
( [ 8 ] ) can be presented in terms of the vector and scalar condensates @xcite@xcite .
vector condensates @xmath45 of the quarks with the flavor @xmath46 ( @xmath47 denotes the ground state of the matter ) are the linear functions of the nucleon densities @xmath2 and @xmath3 . in the asymmetric matter
both su(2 ) symmetric and asymmetric condensates @xmath48 and
@xmath49 obtain nonzero values . in the rest frame of the matter @xmath50 , @xmath51 , @xmath52 .
we can present @xmath53 the values @xmath54 are just the numbers of the valence quarks in the nucleons @xmath55 , @xmath56 , and thus @xmath57 hence , we obtain @xmath58 with @xmath59 the lhs of eq .
( [ 8 ] ) contains the su(2 ) symmetric scalar condensate @xmath60 and su(2 ) asymmetric one @xmath61 these condensates can be presented as @xmath62 @xmath63 is the vacuum value , @xmath64 and @xmath65 the dots in the rhs of eqs .
( [ 15 ] ) and ( [ 16 ] ) denote the terms , which are nonlinear in @xmath66 . in the gas approximation such terms should be omitted .
the su(2 ) invariance of vacuum was assumed in eq .
( [ 16 ] ) .
the expectation value @xmath67 is related to the @xmath68 sigma term @xmath69 , i.e. , @xcite @xmath70 with @xmath71 standing for the current masses of the light quarks , while the value of @xmath69 can be extracted from experimental data on low energy @xmath68 scattering @xcite .
however , the value @xmath72 should be calculated under certain model assumptions on the quark structure of the nucleon .
these were the condensates of dimension @xmath73 .
turning to the condensates of dimension @xmath74 , we find for the gluon condensate @xmath75 @xmath76 is the vacuum value , @xmath77 with the nucleon expectation value @xmath78 obtained in @xcite in a model - independent way .
also , the nonlocal condensate @xmath79 provides the contributions of @xmath74 and those of the higher dimension .
the term of the dimension @xmath74 is @xmath80 with @xmath81 standing for the second moment of the nucleon structure function @xcite .
the condensates of dimensions @xmath82 determine the leading ope terms of the order @xmath83 and @xmath84 of the lhs of eq .
( [ 8 ] ) .
the next to leading terms of the order @xmath85 are determined by the expectation values of the four - quark operators .
the importance of these contributions was analyzed in @xcite . in the gas approximation
they can be presented in terms of the nucleon expectation values .
the latter can be obtained in framework of certain models .
we shall analyze the sum rules in the gas approximation .
it is a reasonable starting point , since the nonlinear contributions to the most important scalar condensate @xmath86 are relatively small at the densities of the order of the phenomenological saturation density @xmath87 of the symmetric matter @xcite .
the qcd sum rules approach , applied to the description of the nucleon self - energies in the symmetric matter @xcite provided the results which are consistent with those , obtained by the methods of nuclear physics . in the case of symmetric matter
@xmath88 , the leading ope terms @xmath89 can be either calculated or expressed in terms of the observables . in the asymmetric case
we need a model for the calculation of the expectation value @xmath72 , defined by eq .
( [ 16 ] ) . in most of the quark models
the nucleon is treated as a system of the valence quarks and the isospin symmetric sea of quark antiquark pairs . under this assumption
the condensate @xmath72 is determined by the valence quarks . in the models with the nonrelativistic valence quarks @xmath90 . in more realistic relativistic models
@xmath91 due to the relativistic reduction .
the calculations of the four - quark condensates require model assumptions on the structure of the nucleon .
the complete set of the four - quark condensates was obtained in @xcite by using the perturbative chiral quark model ( pcqm ) .
the chiral quark model , originally suggested in @xcite , was developed further in @xcite@xcite . in the pcqm
the nucleon is treated as a system of relativistic valence quarks moving in an effective static field .
the valence quarks are supplemented by a perturbative cloud of pseudoscalar mesons , in agreement with the requirements of the chiral symmetry .
this model reflects the main points of the nowadays view on the structure of nucleon .
the latter is treated as the system of the valence quark and the sea - quarks , contained in pions .
the chiral properties are respected . in @xcite the su(2 ) version of pcqm , which includes only the pions ,
have been used .
thus , in the calculations , which include the terms of the order @xmath85 of ope obtained in framework of pcqm we must use the pcqm value @xmath92 @xcite . in the rhs of the sum rules we describe the nucleon by the relativistic in - medium propagator @xcite @xmath93 with the total self - energy @xmath94
we shall use the qcd sum rules for the calculation of the nucleon characteristics @xmath95 identified with the vector self - energy , dirac effective mass and the effective scalar self - energy see , e.g. , @xcite .
two other parameters , to be determined in the approach are the in - medium shifts of the effective values of the nucleon residue @xmath96 and of the continuum threshold @xmath97 .
we present also the result for the single - particle potential energies @xmath98 we trace the dependence of these characteristics on the total density @xmath66 and on the asymmetry parameter @xmath99 eqs .
( [ 1a ] ) and ( [ 2a ] ) .
our approach includes only the strong interactions between the nucleons .
the electromagnetic and weak interactions are neglected . in the asymmetric matter
the characteristics obtain different values for the proton and neutron . considering the asymmetry parameter in the interval @xmath100
, one can see that the values of any parameter @xmath101 for the proton @xmath102 and neutron @xmath103 are connected as @xmath104 we shall present all the values for the proton , using eq .
( [ 25 ] ) to obtain those for the neutron .
we carry out the calculations in the gas approximation , including only terms linear in @xmath66 in the lhs of the sum rules .
thus we can neglect the fermi motion of the nucleons of the matter , which manifest themselves in higher order of density , and put @xmath105 in eq .
( [ 6 ] ) .
having in mind future extensions of the approach , we shall keep the dependence on @xmath106 , using eq .
( [ 26 ] ) for the specific computations . in the simplest hartree approximation ( without multiparticle forces )
@xmath107 is linear in density .
if one neglects the fermi motion of the nucleon , as we did , the same is true for the scalar self - energy @xmath108 .
thus , the parameters @xmath0 and @xmath109 ( [ 23 ] ) exhibit nonlinear behavior due to nonzero value of @xmath110 ( @xmath111 in the mean field approximation ) . in similar way , the characteristics @xmath0 and @xmath109 are not linear in @xmath66 and @xmath99 in our approach due to the relativity large lhs of the sum rules for the structure @xmath112 .
this reflects distribution of the baryon charge between the pole and continuum .
another reason of the nonlinear effects ( although , numerically less important ) is the nonlinear structure of the sum rules equations .
the in - medium nucleon qcd sum rules present nucleon interactions with matter in terms of the exchange by uncorrelated quark - antiquark pairs .
the latter exchanges correspond to the meson exchanges with the same quantum numbers .
thus the contributions of the lowest order ope terms can be viewed to the exchange by the isoscalar and isovector vector and scalar mesons , with the point - like vertices of the interactions , having the standard lorentz structure .
inclusion of the nonlocal effects in the quark condensates corresponds to inclusion of nonlocal effects in the vertices of the interactions between the nucleons of the matter and the mesons .
the four - quark condensates , in which two of the quark operators are averaged over the vacuum state correspond to the anomalous lorentz structures of the vertices of the nucleon - meson interactions . in terms of hadronic degrees of freedom the four - quarks condensates ,
in which all the quark operators act inside the nucleons correspond to the exchanges by the strongly correlated four - quark systems .
such exchanges can be viewed , e.g. , as those by two mesons , interacting with the nucleon at the same point .
this is illustrated by figs . 1 and 2 .
we compare our results for the nucleon self - energies and potential energies with those , obtained by the traditional methods of nuclear physics .
we find a promising consistency of the results .
we recall in brief the main points of qcd sum rules approach in vacuum and in nuclear matter in sec .
ii .
we present the general equations in sec .
iii , providing the solutions in sec .
iv .
we compare the results to those , obtained by the methods of nuclear physics in sec . v. we summarize in sec .
vi .
* i. general equations * * a. sum rules in vacuum * to make the paper self - consistent , we recall the main points of the qcd sum approach in vacuum @xcite . the function @xmath19 ( often referred to as `` polarization operator '' ) is presented as @xmath113 with @xmath114 being the three - quark local operator ( often referred to as `` current '' ) with the proton quantum numbers .
the usual choice is @xcite @xmath115 \gamma_5\gamma_\mu d^c(x)\ , \label{28}\ ] ] where @xmath116 denotes a transpose and @xmath117 is the charge conjugation matrix .
the upper indices denote the colors . the lhs of eq .
( [ 4 ] ) is approximated by several lowest terms of ope , i.e. , @xmath118 .
the empirical data are used for the spectral function @xmath119 on the rhs of eq .
( [ 4 ] ) .
namely , it is known , that the lowest laying state is the bound state of three quarks , which manifests itself as a pole in the ( unknown ) point @xmath120 . since the next singularity is the branching point @xmath121 , one can present @xmath122 with @xmath123 being the residue at the nucleon pole . here
@xmath15 correspond to the two lorentz structures of eq .
( [ 3 ] ) , @xmath124 , @xmath125 . thus , eq . (
[ 4 ] ) takes the form @xmath126 of course , the detailed structure of the spectral density @xmath127 can not be resolved in such approach .
the further approximations are based on the asymptotic behavior @xmath128 at @xmath129 with @xmath130 denoting the discontinuity .
the discontinuity is caused by the logarithmic contributions of the perturbative ope terms .
the usual anzats consists in extrapolation of eq .
( [ 31 ] ) to all the values of @xmath131 , replacing also the physical threshold @xmath132 by the unknown effective threshold @xmath27 , i.e. , @xmath133 thus eq . (
[ 4 ] ) takes the form @xmath134 the lhs of eq .
( [ 33 ] ) contains qcd condensates .
the rhs of eq .
( [ 33 ] ) contains three unknown parameters : @xmath135 and @xmath27 .
the ope becomes increasingly true while the value @xmath22 increases .
the `` pole + continuum '' model is more accurate at the smaller values of @xmath22
. thus one can expect eq .
( [ 33 ] ) to be true in a certain limited interval of the values of @xmath22 . to improve the overlap of ope and phenomenological descriptions one usually applies the borel transform defined as @xmath136 with @xmath137 called the borel mass .
it is important in the applications to the sum rules that borel transform eliminates the polynomials and emphasizes the contribution of the lowest state in rhs of eq .
( [ 33 ] ) due to relation @xmath138 the borel - transformed form of eq .
( [ 33 ] ) reads @xmath139 and is known as qcd sum rules . actually , there are two sum rules for the structures @xmath140 and @xmath141 of the function @xmath40 defined by eq .
( [ 3 ] ) .
it appeared to be more convenient to work with eq .
( [ 36 ] ) multiplied by the numerical factor @xmath142 .
the two sum rules for the nucleon in vacuum can be presented in the form @xcite @xmath143 with @xmath144 here @xmath145 , @xmath146 the lhs of eqs .
( [ 37 ] ) and ( [ 38 ] ) @xcite have been obtained with the account of the condensates of dimension @xmath147 , i.e. , with the account of the terms of the order @xmath148 in ope of the functions @xmath149 @xmath150 with the traditional notations @xmath151 ( we assumed the isotopic invariance of vacuum @xmath152 ) , @xmath153 , @xmath154 . here
@xmath155 are the functions of the ratio @xmath156 : @xmath157 .
they are determined by the formulas @xmath158 the factor @xmath159 accounts for the anomalous dimension , i.e. , the most important corrections of the order @xmath18 enhanced by the `` large logarithms '' . in eq .
( [ 44 ] ) @xmath160gev , while @xmath161gev is the normalization point of the characteristic involved .
note that the two last terms on the rhs of eq .
( [ 41 ] ) originate from the four - quark condensates @xmath162 and can be expressed through the single term @xmath163 only in framework of factorization hypothesis @xcite@xcite . also , the last term on the rhs of eq .
( [ 42 ] ) is the six - quark condensate , evaluated in the same approximation .
the matching of the lhs and rhs of eqs .
( [ 37 ] ) and ( [ 38 ] ) have been achieved @xcite in the domain @xmath164 providing the values of the vacuum parameters @xmath165 if @xmath166 gev .
* b. sum rules in nuclear matter * the ope terms of the polarization operator in nuclear matter @xmath167 contains the in - medium values of qcd condensates .
some of these condensates vanish in vacuum , obtaining non - zero values only in medium .
the other ones just change their values compared to the vacuum ones .
the spectrum of the function @xmath28 is much more complicated , than that of the vacuum function @xmath19 .
however , @xcite@xcite , the spectrum of the function @xmath168 at fixed value of @xmath106 can be described by the `` pole + continuum '' model at least until we include the terms of the order @xmath169 in the ope of @xmath168 .
using eqs .
( [ 21 ] ) - ( [ 23 ] ) , we present the nucleon propagator as @xmath170 with @xmath0 and @xmath109 defined by eq .
( [ 23 ] ) , while the new position of the nucleon pole is @xmath171 and @xmath172 thus , we shall present the dispersion relations for the functions @xmath37 @xmath38 in the form @xmath173 with @xmath174 , @xmath175 , @xmath176 .
the borel - transformed sum rules take the form @xmath177 with @xmath178 here @xmath179 is the effective value of the nucleon residue in nuclear matter .
actually , we shall solve the sum rules equations , subtracting the vacuum effects @xmath180 with @xmath181 .
the functions @xmath182 , @xmath183 and @xmath184 are defined by eqs .
( [ 39 ] ) , ( [ 41 ] ) and ( [ 42 ] ) , while @xmath185 .
* iii .
contributions to the left - hand side of the sum rules * we shall use the current , presented by eq .
( [ 28 ] ) for the calculation of the function @xmath28 , defined by eq .
( [ 47 ] ) .
the possibility of using it in isospin asymmetric system requires additional argumentation , which is given in appendix a. we shall include the terms of the order @xmath83 , @xmath84 and @xmath85 in the lhs of the sum rules ( [ 51 ] ) .
this corresponds to the terms of the order @xmath186 , @xmath187 and @xmath188 in the lhs of the borel transformed equations ( [ 57])([59 ] ) .
we shall include subsequently the contributions of three types .
the terms @xmath189 stand for the lowest order local condensates .
they contribute to @xmath83 and @xmath84 terms of eq .
( [ 51 ] ) .
these contributions are similar to simple exchanges by isovector vector and scalar mesons between the nucleon on the rhs of eq .
( [ 51 ] ) and the nucleons of the matter , figs .
1(a),2(a ) .
the terms @xmath190 are caused by the nonlocalities of the vector condensate , fig .
1(d ) , contributing to the terms of the order @xmath84 and @xmath85 .
they correspond to the account of the form factors in the vertices between the isovector mesons couple to the nucleons , fig .
2(d ) .
finally , @xmath191 describes the contributions of the four - quark condensates , figs .
1(b),1(c ) , being of the order @xmath85 .
they correspond to the two - meson exchanges ( or to exchanges by four - quark mesons , if there are any ) and to somewhat more complicated structure of the meson - nucleon vertices , fig .
2(c ) .
thus we present the lhs of eqs .
( [ 52])([54 ] ) as @xmath192 and the lhs of eqs .
( [ 57])([59 ] ) @xmath193 with @xmath194 , @xmath195 , @xmath196 , while @xmath197 and @xmath198 are the corresponding contributions in the vacuum case .
the calculation of the function @xmath168 defined by eq .
( [ 47 ] ) is based on the presentation of the single - quark propagator in medium @xmath199 with @xmath200 being the free propagator of the quark in the chiral limit .
in the ope calculations we neglect the values of the current masses of the light quarks since they are small in our scale . recall that @xmath46 denotes the light quark flavor . in the lowest orders of ope two of the quarks
are described by the free propagators and only one of the quarks is presented by the second or the third term of the rhs of eq .
( [ 62 ] ) .
inclusion of the higher order ope terms will require the two - quark in - medium propagators . at @xmath201
the matrix elements in the second and third terms on the rhs are just the vector and scalar condensates defined by eqs .
( [ 9 ] ) and ( [ 15 ] ) . the contribution of the bilocal configurations can be expressed in terms of the higher moments and twists of the nucleon structure functions @xcite . the bilocal operators on the rhs of eq .
( [ 62 ] ) are not gauge invariant .
the gauge invariant expression , achieved by substitution @xcite @xmath202 with @xmath203 standing for the covariant derivatives , provides the infinite set of the local condensates .
the expectation values in eq .
( [ 62 ] ) depend on the variables @xmath204 and @xmath205 . in the gas approximation
we need only the nucleon matrix elements . for the vector structure
the general form is @xmath206 with @xmath207 defined by eq .
( [ 63 ] ) .
the contribution of the vector structure ( [ 64 ] ) to the polarization operator @xmath28 can be presented as @xmath208 here we denoted @xmath209 for any 4-vector @xmath210 . for the scalar structure
@xmath211 * a. local condensates of the lowest dimensions * to include the local vector and scalar quark condensates of the lowest dimensions , we must put @xmath201 in eq .
( [ 64 ] ) and in the matrix element @xmath212 in eq .
( [ 67 ] ) .
thus @xmath213 while @xmath214 we must add also the contribution of the gluon condensate .
technically it can be obtained by replacing the vacuum gluon condensate by it s in - medium value eqs .
( [ 18 ] ) and ( [ 19 ] )
. the contributions of these local condensates can be expressed as @xmath215 with the dependence on @xmath66 and @xmath99 being contained in the factors @xmath216 and the function @xmath217 , given by eqs .
( [ 18 ] ) and ( [ 19 ] ) .
the other functions are @xcite @xmath218 the notation @xmath219 , eq .
( [ 43 ] ) , means that the functions depend on the ratio @xmath220 .
now we must find the @xmath99-dependence of the nucleon self - energies and also of the parameters @xmath221 and @xmath222 .
note that there is a simple solution of eqs .
( [ 52 ] ) and ( [ 54 ] ) @xmath223 which is true with the same accuracy as the solutions for the symmetric matter @xmath88 @xcite .
indeed , assuming that @xmath221 does not change with @xmath99 ( [ 73 ] ) , we find that the function @xmath224 in the lhs of eq .
( [ 52 ] ) should not depend on @xmath99 with the accuracy of @xmath225 . ] .
this leads to eqs .
( [ 74 ] ) and ( [ 75 ] ) .
* b. inclusion of the nonlocal condensates * now we present the functions in the rhs of eq .
( [ 64 ] ) as @xmath226 due to su(2 ) invariance they can be presented in terms of the proton functions @xmath227 @xmath228 with @xmath229 .
expansion in powers of @xmath205 corresponds to the expansion of the function @xmath28 in powers of @xmath13 . to obtain the terms of the order @xmath17 it
is sufficient to include two lowest terms of the expansions in powers of @xmath205 .
one can present @xcite @xmath230 with @xmath231 here @xmath232 is the contribution of the quarks with the flavor @xmath46 to the asymptotic of the nucleon structure function @xmath233 , normalized by the condition @xmath234 with the rhs presenting just the number of the valence quarks in the nucleon .
thus , expansion of the function @xmath235 in powers of @xmath204 is expressed through the moments of the distributions @xmath236 .
the moments are well known at least , those , which are numerically important .
also , the first moments @xmath237 of the distribution @xmath238 were calculated in @xcite by the qcd sum rules method : @xmath239 the moments of the function @xmath240 can be obtained by using the equations of motion @xmath241 .
thus , in the chiral limit @xcite @xmath242 here we denoted @xmath243 for any function @xmath244 .
note that the nonlocality of the scalar condensate , i.e. , of the last term on the rhs of eq .
( [ 62 ] ) does not manifest itself in the terms up to @xmath85 .
the first derivative in @xmath204 , as well all the derivatives of the odd order vanish in the chiral limit due to qcd equation of motion .
the next to leading order of the expansion in powers of @xmath205 vanishes due to certain cancellations @xcite as well as in the case of vacuum @xcite for the particular choice of the operator @xmath245 presented by eq .
( [ 28 ] ) .
we neglect the terms , containing more complicated gluon condensates , since they are numerically small @xcite . using eqs .
( [ 64 ] ) and ( [ 77 ] ) we obtain for the vector structure ( [ 65 ] ) @xmath246 the further calculations , employing eqs .
( [ 78 ] ) and ( [ 79 ] ) are similar to those , carried out for the case of symmetric matter @xcite see appendix b. finally , the higher moments and higher twists of the nucleon structure functions provide the contributions @xmath247 to the lhs @xmath248 of the sum rules eq . ( [ 60 ] ) @xmath249\ , \nn\\ & & \quad u^q_{n,2}(m^2)=\frac{8\pi^2}{3l^{4/9}m}\left [ \frac32\,m^2m^2e_{0m}(\langle\eta^u\alpha\rangle -\langle\eta^d\alpha\rangle ) \right ] ; \label{85 } \\
\nn\\ \nn\\ & & u^p(m^2)\ = \ ( u^p_{n,1}(m^2 ) + \beta u^p_{n,2}(m^2))\rho\ , \nn\\ & & \quad u^p_{n,1}(m^2)=\frac{8\pi^2}{3l^{4/9}}\bigg[-5\left(m^4 e_{1 m } -(s - m^2)m^2 e_{0m}\right)\langle\eta\alpha\rangle \nn\\ & & \quad -\ \frac{12}5 m^2m^2e_{0m}\langle\eta\alpha^2\rangle + \ \frac{18}5m^2m^2e_{0m}\langle\xi\rangle\bigg]\ , \nn \\ & & \quad u^p_{n,2}(m^2)=\frac{8\pi^2}{3l^{4/9}}\bigg [ 3\left(m^4e_{1m}-(s - m^2)m^2e_{0m}\right ) \langle(\eta^u-\eta^d)\alpha\rangle
\nn\\ \label{86 m } & & \quad+\ \frac95\,m^2m^2e_{0m}\langle(\eta^u-\eta^d)\alpha^2\rangle -\frac{27}{10}\,m^2m^2e_{0 m } \langle(\xi^u-\xi^d)\rangle \bigg ] ; \\ \nn\\ \nn\\ & & u^i(m^2)\ = \ 0\ . \label{85a}\end{aligned}\ ] ] here we denote @xmath250 , @xmath251 , @xmath252 , eq .
( [ 43 ] ) , the value of @xmath253 is determined by eq .
( [ 81 ] ) . * c. inclusion of the four - quark condensates * the exchange by two quark - antiquark pairs between the current ( [ 28 ] ) and the matter is described in terms of the four - quark expectation values @xmath254 with @xmath255 being the basic @xmath12 matrices @xmath256 acting on the lorentz indices of the quark operators .
we did not display the color indices in eq .
( [ 86 ] ) , keeping in mind that the quark operators are color antisymmetric . in the gas approximation @xmath257 the characteristics @xmath258 and @xmath259 can be presented as @xmath260 the first two ( `` factorized '' ) terms in the rhs of eqs .
( [ 89 ] ) and ( [ 90 ] ) describe two quark operators , acting on the vacuum state , while the other two operators act inside the nucleon .
the last terms describe the `` internal '' action of all four operators inside the nucleon .
this is shown by the lower index @xmath261 .
the coefficients 5/6 and 2/3 on rhs of eqs .
( [ 89 ] ) and ( [ 90 ] ) present the weights of the color - antisymmetric states @xcite .
the contribution of the four - quark expectation values to the in - medium change of the polarization operator can be written as @xmath262 the matrices @xmath263 and @xmath264 can be obtained by using the general expression for the function @xmath28 presented in @xcite @xmath265 here @xmath266 if @xmath267 has a vector or tensor structure , while @xmath268 in the scalar , pseudoscalar and axial cases .
the sign is determined by that of the commutator between matrix @xmath267 and the charge conjugation matrix @xmath117 eq .
( 16 ) .
the products @xmath269 obtain nonzero values if the matrices @xmath270 and @xmath267 have the same lorentz structure . in this case all the structures presented by eq .
( [ 87 ] ) contribute to @xmath271 .
the products @xmath272 do not turn to zero only if @xmath267 has a vector or axial structure . in the latter case @xmath270
should be an axial matrix as well . in the former case @xmath270
can be either lorentz scalar or lorentz vector .
we denote @xmath273 , @xmath274 and @xmath275 , @xmath276 for the similar lorentz structures @xmath277 and @xmath278 .
the scalar and pseudoscalar expectation values are lorentz scalars .
thus , their contributions can be expressed through single parameters @xmath279 , @xmath280 , @xmath281 .
the latter is true also for the scalar - vector expectation value @xmath282 .
we obtain @xmath283 in the other channels the four - quark condensates have more complicated structure . in the vector and axial channels
@xmath284 in the tensor channel @xmath285 with @xmath286 using eq .
( [ 92 ] ) we obtain @xmath287 we can write eq .
( [ 91 ] ) in the form @xmath288 with each of the coefficients @xmath289 being the composition of the contributions of the neutrons and protons of the matter @xmath290 and @xmath291 , i.e. , @xmath292 using eqs .
( [ 93 ] ) ( [ 95 ] ) we obtain @xmath293 here we denoted @xmath294 , @xmath295 .
recall that the lower index @xmath296 stands for the neutron or the proton . in the last equality we defined @xmath297 , neglecting the nonlocal contributions to the scalar - vector condensate .
however in the scalar channel such terms provide the contributions of the order @xmath85 to the rhs of eq .
( [ 98 ] )
. they will be included below appendix c. the calculations of the contributions to rhs of eq .
( [ 100 ] ) require the model assumption on the structure of the nucleon .
the complete set of the four - quark condensate was obtained in @xcite by using the perturbative chiral quark model ( pcqm ) .
the chiral quark model , originally suggested in @xcite , was developed further in @xcite@xcite . in the pcqm
the nucleon is treated as a system of relativistic valence quarks moving in an effective static field .
the valence quarks are supplemented by a perturbative cloud of the pseudoscalar mesons in agreement with the requirements of the chiral symmetry . in @xcite the simplest su(2 )
version of pcqm , which includes only the pions , have been used .
the parameters which enter eq .
( [ 100 ] ) are calculated in this version of pcqm .
there are three types of contributions to the four - quark condensate in the framework of this approach .
all four operators can act on the valence quarks . also , four operators can act on the pion .
there is also a possibility that two of the operators act on the valence quarks while the other two act on the pions .
following @xcite we speak of the `` interference terms '' in the latter case . to obtain the contribution of the pion cloud , we need the expectation values of the four - quark operators in pions .
the latter have been deduced in @xcite by using the current algebra technique .
it was shown in @xcite , that in the case of the symmetric matter there is a remarkable cancellation of the pion contributions in the function @xmath28 . since the pion contents of the asymmetric matter manifests itself in the different intensities of @xmath298 and @xmath299 fields , while the four - quark expectation values in @xmath298 and @xmath299 mesons are the same , the cancellation takes place in this case as well .
this cancellation takes place in any model of the nucleon which treats the pion cloud perturbatively .
thus , the contributions of the four - quark condensates come from the terms , determined by the valence quarks only and from the interference terms .
this enables us to present eq .
( [ 89 ] ) as @xmath300 corresponds to the sum of the term in which all the four operators act on the valence quarks and the term in which two of the operators act on the valence quarks while the other two act on pions . of course , the first ( `` factorized '' ) term in the rhs of eq .
( [ 101 ] ) obtains a nonvanishing value only in the scalar case @xmath301 .
the expectation values of the operators of different flavors , providing nonvanishing contributions to the rhs of eq .
( [ 90 ] ) are the scalar - vector condensate @xmath302 and @xmath303 with @xmath277 standing for vector or axial structures .
the meaning of the lower index @xmath304 is the same as in eq .
( [ 101 ] ) .
note that among the interference terms contributing to the four - quark condensates , there is so - called `` vertex interference '' , in which one of the pcqm vertices of the self - energy of the valence quark is replaced by the four - quark operator .
some of such terms contain the matrix elements @xmath305 and @xmath306 , contributing to the expectation values @xmath307 , being connected with the matrix elements @xmath308 of all structures @xmath270 by the fierz transform . on the other hand , they depend on the values of the quark masses , since @xmath309 with @xmath310 and @xmath311 denoting the mass and the decay constant of pion . in a somewhat straightforward approach one
substitutes the current quark masses . following more sophisticated models of the pions @xcite one should substitute the constituent quark masses , thus obtaining the values , which are negligibly small in our scale . using the complete set of the nucleon four - quark expectation values @xcite , we obtain @xmath312 with the coefficients @xmath313 being determined by eqs .
( [ 101])([103 ] ) , while @xmath314 we use the value @xmath315=(-241 mev)@xmath316 , corresponding to @xmath317=0.55 gev@xmath316 , employed in @xcite .
note that @xmath317 is just a convenient scale for presentation of the results .
it does not reflect the chiral properties of @xmath318 . the @xmath319 term results mainly as the sum of the expectation value of the product of the four @xmath320-quark operators , described by the first ( `` factorized '' ) term on the rhs of eq .
( [ 101 ] ) , and that of the product of two @xmath320 and two @xmath321-quark operators in the vector channel eq . (
[ 103 ] ) .
the former contributions depend on @xmath99 , while the latter do not .
thus , the coefficient @xmath322 depends on @xmath99 strongly .
the @xmath323 term is determined mostly by the expectation value ( [ 103 ] ) in the vector channel , with the protons and neutrons providing the equal contributions .
this explains the weak dependence of the parameter @xmath324 on @xmath99 .
the coefficient @xmath325 is dominated by the first term on the rhs of eq .
( [ 102 ] ) , providing stronger dependence on @xmath99 .
the calculations give @xmath326 in the simplified model of the pion , which does not include the renormalization of the quark masses by the interactions , the value of the coefficient @xmath322 is somewhat different @xmath327 while the values of @xmath324 and @xmath325 remain unchanged .
the contributions of the four - quark condensates to the lhs of the borel transformed sum rules ( [ 57])([59 ] ) can be presented in the same way , as for the symmetric case @xmath328 * iv .
solutions of the sum rules equations * now we present the solutions of the sum rules equations , focusing on the functions @xmath329 and @xmath330
. we shall include the terms @xmath331 and @xmath332 ( eq . ( [ 61 ] ) ) in the succession in lhs of eqs .
( [ 57])([59 ] ) .
two lowest order ope contributions to the vector structures @xmath319 and @xmath323 are presented in terms of the vector and gluon condensates and of the nucleon structure functions .
these characteristics are either calculated in a model - independent way , or determined in the experiments .
the lowest order ope terms in the scalar channel are expressed in terms of isotope symmetric and isotope asymmetric scalar condensates @xmath333 and @xmath334 eqs .
( [ 15 ] ) and ( [ 16 ] ) . here
the situation becomes somewhat more complicated .
the expectation value @xmath67 is related to the @xmath68 sigma - term @xmath69 by eq .
( [ 17 ] ) .
the value of @xmath69 can be extracted from the data on low - energy @xmath68 scattering .
the procedure consists in subtracting the high - order chirality - violating terms @xmath335 from the experimental value @xmath336 , i.e. , @xmath337 .
the value @xmath338mev was obtained in @xcite by dispersion relation technique . however , there are some uncertainties in deducing the value of @xmath336 from the experimental data .
the canonical value @xmath339mev @xcite is now challenged by the higher values @xmath340mev @xcite .
assuming @xmath341ev @xcite , we find that @xmath342mev corresponds to @xmath343 .
additional uncertainties emerge , since the true value of the sum @xmath344 may be somewhat larger .
there is no experimental data on the expectation value @xmath72 . if the nucleon is treated as a system of the valence quarks and the isospin symmetric sea of the quark
antiquark pairs , the expectation value @xmath72 is determined by the contribution of the valence quarks .
thus , the reasonable values are @xmath90 for the nonrelativistic models and @xmath91 in the relativistic case . until we include only the leading ope terms @xmath345 , we can solve the sum rules equations for any values of @xmath67 and @xmath72
. however the four - quark condensates are obtained in framework of a specific perturbative chiral quark model ( pcqm ) . within this model
@xmath346mev @xcite , leading to @xmath343 .
thus , to be self - consistent , we must use this value as the basic one in the general equations , which include the contributions @xmath332 .
note also , that the values of @xmath336 extracted from the experimental data are correlated with the assumption on the strange quark content @xmath347 @xcite .
the values @xmath348mev correspond to @xmath349 , with a large part of the nucleon mass being due to the strange quarks .
the smaller values of @xmath336 require much smaller values of @xmath350
. in pcqm one finds @xmath351 @xcite , in agreement with the smaller values of @xmath336 .
the pcqm value @xmath92 can be obtained by using the results of @xcite .
we shall present most of the numerical results for the values @xmath352 anyway , in the linearized version of the sum rules , the nucleon characteristics will be presented as the explicit function of the condensates , e.g. , of the parameters @xmath67 and @xmath72 .
* a. solution of the general equations * here we present the solutions of the general equations ( [ 57])([59 ] ) , which are identical to eqs . ( [ 52])([54 ] ) . recall that we approximate the in - medium condensates by the functions , which are linear both in @xmath66 and @xmath99
. however the solutions @xmath329 and @xmath330 are not linear .
one can demonstrate this by presenting eqs .
( [ 53 ] ) and ( [ 54 ] ) as @xmath353 with the density and @xmath99 dependence of @xmath354 leading to nonlinear behavior of @xmath0 and @xmath109 ( even if we assume @xmath355 )
. the nonlinear dependence of rhs of eq .
( [ 110 ] ) on @xmath221 also cause the nonlinear contributions to @xmath0 and @xmath109 .
however , they are numerically less important . now we include the terms @xmath331 and @xmath332 in succession in lhs of eqs .
( [ 57])([59 ] ) . * _ 1 .
the role of the lowest order local condensates _ * as we have seen , these are the contributions , which contain the vector condensate @xmath356 , gluon condensate @xmath217 and the scalar condensates @xmath357 and @xmath358 eqs .
( [ 11 ] ) , ( [ 13 ] ) , ( [ 14 ] ) , ( [ 18 ] ) . account of these terms corresponds to the one - meson exchanges between the nucleon under consideration and the nucleon of the matter , with the point - like structures of the meson
nucleon vertices , fig .
2(a ) .
the solution can be obtained by using eqs .
( [ 68])([70 ] ) for the functions @xmath359 .
as we have seen , there is a simple solution , expressed by eqs .
( [ 73])([75 ] ) .
the procedure of minimization of the difference between the lhs and rhs of eqs .
( [ 57])([59 ] ) indeed prefers the values @xmath360 .
thus eqs .
( [ 73 ] ) and ( [ 74 ] ) appear to be true with good accuracy .
hence , @xmath361 thus , in the matter with the excess of the neutrons @xmath362 , we obtain @xmath363 , and @xmath364 .
for example , using the value @xmath365 obtained by sum rules approach in @xcite ( @xmath366 mev ) , we find @xmath367mev , @xmath368mev for the neutron matter @xmath369 at @xmath370 .
the minimization procedure choses @xmath371 , @xmath372 , @xmath373 , @xmath367mev and @xmath374mev .
thus , eqs .
( [ 111 ] ) work well indeed . however , this approximation is not sufficient for the description of the potential energy @xmath375 ( [ 24 ] ) , providing @xmath376 for both symmetric and asymmetric cases . *
_ 2 .
the role of the four - quark condensates _
* now we include the four - quark condensate , i.e. , we use eq .
( [ 60 ] ) for @xmath377 , with @xmath332 described by eqs .
( [ 108 ] ) .
inclusion of these terms mimics several contributions on the rhs of the sum rules ( [ 57])([58 ] ) . in the condensates @xmath378 , presented by eq .
( [ 101 ] ) the first term , which has a nonvanishing value only if @xmath301 , generates a contribution to the @xmath319 structure due to the anomalous lorentz structure of the interaction between the scalar field and the nucleon , caused by the chiral - odd vacuum condensate @xmath379 . in similar way the first term on the rhs of eq .
( [ 102 ] ) describes the contribution of the vector meson exchange to the scalar structure of the nucleon propagator .
the anomalous lorentz structures emerge if the nucleon - meson vertices are treated beyond the lowest order .
these contributions are illustrated by fig .
2(b ) .
the second terms on the rhs of eqs .
( [ 101 ] ) and ( [ 102 ] ) and on the rhs of eq .
( [ 103 ] ) describe exchanges by the four - quark strongly correlated system ( see fig .
2(c ) ) .
the condensates presented by eq .
( [ 103 ] ) have the same values for the proton and neutron .
thus their contributions do not depend on @xmath99 .
the differences between the neutron and proton characteristics in the neutron matter at @xmath370 become @xmath380mev , @xmath381mev . in the simplified model for the pion with the current masses of the constituent quarks , where @xmath382 is given by eq .
( [ 107 ] ) , we obtain @xmath383 mev , while @xmath384 mev at these values of @xmath66 and @xmath99 .
* _ 3 .
the final results _ * these contributions come from the account of the @xmath385-dependence of the expectation values of the vector operators @xmath386 with @xmath207 defined by eq .
( [ 63 ] ) .
as we have seen , the nonlocality of the scalar condensates is not important in our case . the nonlocality is included by putting @xmath387 with @xmath388 defined by eq .
( [ 85 ] ) .
we use the structure functions , obtained in @xcite for the calculation of the terms @xmath389 and @xmath390 .
account of the nonlocality of the vector condensate corresponds to inclusion of the formfactor of the vertex of the interaction between the vector meson and the nucleon of the matter - fig .
2(d ) .
recall that similar contribution for the effective scalar meson exchanges vanishes in our approximation - sec .
iii.b .
we find the dependence of the nucleon vector self - energy @xmath0 and of the effective mass @xmath109 on density of the matter and on the asymmetry parameter @xmath99 and show the results in figs . 3 and 4 .
for example , the differences between the neutron and proton characteristics in the neutron matter at @xmath391 are @xmath392 mev , @xmath393 mev . in the simplified model for the pion with the current quark masses , where @xmath382 is given by eq .
( [ 107 ] ) we obtain @xmath394 mev , @xmath395 mev at these values of @xmath66 and @xmath99 .
the minor change is due to the small change in the @xmath99 dependence of the condensate @xmath322 . due to the nonlinear character of eqs .
( [ 57])-([59 ] ) the self - energies @xmath329 and @xmath396 could have been nonlinear in both @xmath66 and @xmath99 .
the nonlinear behavior of these characteristics with @xmath66 manifests itself explicitly .
however , the dependence on @xmath99 appears to be linear in framework of the accuracy of our computations ( see the next subsection ) .
thus both @xmath0 and @xmath1 can be approximated by linear functions of @xmath99 @xmath397 with @xmath398 for the proton , @xmath399 for the neutron .
the functions @xmath400 and @xmath401 are shown in fig .
5 .
they can be approximated by polynomials of the second order see appendix d. the single - particle potential energy is expressed by eq .
( [ 24 ] ) . at @xmath370
the neutron - proton difference of the potential energy caused by the isovector interaction is @xmath402 mev at small @xmath99 . in fig .
6 we show the dependence @xmath403 for several values of @xmath99 for both neutrons and protons .
recall that the potential energy is determined with the lower accuracy then the self - energies .
the nucleon residue @xmath404 and the spectrum threshold @xmath221 exhibit very weak dependence on @xmath99 .
thus we can assume @xmath405 * b. explicit expression for the nucleon parameters in terms of qcd condensates * we can present an approximate solution of eqs .
( [ 57])([59 ] ) , in which the nucleon self - energies are expressed in terms of the qcd condensates explicitly .
we see from eq .
( [ 116 ] ) , that @xmath406 , while @xmath407 is close to it s vacuum value @xmath27 eq . ( [ 46])@xcite .
thus , we can put @xmath355 in the rhs of eqs .
( [ 52])-([54 ] ) and ( [ 110 ] ) .
this enables us to present the proton characteristics @xmath0 and @xmath109 as the explicit functions of the quark condensates and of the borel mass @xmath187 : @xmath408 \frac{1}{1+f^q(\rho,\beta ) } \label{118}\end{aligned}\ ] ] with @xmath409 \frac\rho{\rho_0}\ .\ ] ] we denote @xmath410 and @xmath411 for other @xmath114 and introduce for @xmath412 @xmath413 with the functions @xmath414 and @xmath415 defined by eqs .
( [ 72 ] ) and ( [ 108 ] ) .
it is instructive to present the density @xmath66 in units of the observable saturation density of the symmetric matter @xmath87 .
note that the functions @xmath416 defined by eq .
( [ 120 ] ) @xmath417 depend on @xmath187 rather weakly .
thus , approximating @xmath418 we can replace the functions @xmath416 in the lhs of eqs .
( [ 120 ] ) by the constant coefficients @xmath419 .
numerically most important functions @xmath420 and @xmath421 can be approximated by the constant values with the errors of about 4% and 7% .
the largest errors of about 25% emerge in the averaging of the functions @xmath422 .
this solves the problem of expressing the in - medium change of nucleon parameters through the values of the condensates . for the proton @xmath423 @xmath424\frac1{(1-{\cal f}_q)}\ , \ ] ] with @xmath425 the coefficients in the rhs of eqs .
( [ 127a])([128 ] ) are @xmath426 note that the dependence of @xmath427 on @xmath99 is very weak .
( recall that the lowest order ope terms in the @xmath319 structure of polarization operator do not depend on @xmath99 . )
thus , in the lhs of eqs .
( [ 127a ] ) , ( [ 127 ] ) only the dependence of the numerators on @xmath99 is important .
this explains the linear dependence of the self - energies @xmath0 and @xmath1 on @xmath99 .
the values of @xmath0 , @xmath109 and @xmath427 for the neutron are described by eqs .
( [ 127a])([128 ] ) with @xmath99 changed to @xmath428 .
equations ( [ 127a ] ) , ( [ 127 ] ) enable to obtain @xmath0 and @xmath109 during the successive inclusion of condensates of the higher dimensions . if only the leading ope terms are included the values provided by eqs .
( [ 127a ] ) , ( [ 127 ] ) actually coincide with the solutions of eqs .
( [ 57])([59 ] ) .
if all the contributions are included eqs .
( [ 127a ] ) , ( [ 127 ] ) reproduce the values of @xmath0 and @xmath109 with the accuracy of 15@xmath429 and 10@xmath429 correspondingly for the symmetric matter .
the precision of eqs .
( [ 127a ] ) and ( [ 127 ] ) changes with @xmath99 . for the neutron matter
these equations provide the values @xmath430 mev and @xmath431 mev , comparing to the values of 110 mev and -70 mev , obtained in the previous subsection from the general solution of eqs .
( [ 57])([59 ] ) . * v. discussion * now we compare our results to those , obtained by nuclear physics methods .
the lowest order ope terms in the lhs of the sum rules describe mainly the exchanges by the localized @xmath432 pairs .
this corresponds to the vector and ( effective ) scalar meson exchanges between the nucleon and the nucleons of matter .
these exchanges have the point - like vertices and the standard lorentz structures .
inclusion of the higher order ope terms corresponds to a more complicated picture of the meson exchanges between the nucleons . turning to the four - quark condensates , we can separate the two types of terms . in the `` factorized '' contributions one of the @xmath432 operators
is averaged over vacuum . in the `` internal '' terms both @xmath432 pairs act inside the nucleons , eqs .
( [ 89 ] ) , ( [ 90 ] ) , ( [ 101])-([103 ] ) .
the first ( factorized ) term in the rhs of eq .
( [ 101 ] ) describes the contribution to the @xmath433structures of the polarization operator , which contains the scalar expectation value @xmath434 . in similar
way the first ( factorized ) term in the rhs of eq .
( [ 102 ] ) contributes to the scalar structure of the polarization operator , being proportional to the vector expectation value @xmath435 .
these terms correspond to the anomalous lorentz structures of the nucleon - meson vertices . as to the `` internal '' terms , i.e. , the last terms on the rhs of eqs .
( [ 89 ] ) , ( [ 90 ] ) , they can be interpreted as the exchanges by two - meson systems with their local interactions with the nucleon or as the exchanges by four - quark mesons ( if there are any ) .
inclusion of the nonlocal vector condensates @xmath436 means that the vertices of the interactions between the nucleons of the matter and the vector mesons do not have a point - like structure , requiring rather description by the formfactors .
the nonlocality of the scalar condensate does not influence the results in our approach .
a usual subject of calculation is the difference between the characteristics of neutron and proton .
if only the lowest ope terms are included , the vector self - energies are determined by the vector condensates .
the neutron - proton difference @xmath437 , usually attributed to the @xmath66 meson exchange is 170 mev at @xmath370 and @xmath438 .
inclusion of the four - quark condensates and of the nonlocalities subtract 30 mev and 28 mev from this value .
the lowest order ope terms provide the difference of the effective masses @xmath439 = 50 mev at the same values of @xmath66 and @xmath99 .
inclusion of the four - quark condensates and of the nonlocalities adds ( -160 mev ) and 40 mev .
this leads to @xmath440 mev and @xmath439 = -70 mev in neutron matter at @xmath370 the structure of the equations for @xmath0 amd @xmath109 ( [ 127a ] ) , ( [ 127 ] ) is similar to that ( [ 23 ] ) employed in nuclear physics . recall that in the hartree approximation dependence of @xmath0 on the density deviates from the linear law due to @xmath110 ( [ 23 ] ) .
( the same refers to @xmath1 if the nucleon fermi motion is neglected ) . in our approach
the nonlinear behavior of the self energies is due to nonzero values of @xmath427 .
now we compare the numerical results . considering the papers , containing relativistic calculations , we can compare the vector and scalar self - energies @xmath0 and @xmath441 . in the case of the works ,
carried out in the nonrelativistic approximation , we can compare the nucleon potential energies @xmath442 .
we analyze also the contribution to the parameter , conventionally denoted as @xmath443 @xcite , which is defined as @xmath444 being thus the lowest order term of @xmath445 expansion of the averaged binding energy @xmath446 per nucleon at saturation value of density . of course
, we can not expect very good agreement , since our calculations are carried out in the gas approximation .
the future and more sophisticated calculations should include the scalar and four - quark condensates beyond the gas approximation .
this would correspond to the account of the renormalization of the nucleon interactions with the matter by the particle - hole excitations in rhs of the sum rules .
another reason is that the results should be corrected for the effects of antisymmetrization of the total final state wave function ( `` exclusion effect '' ) @xcite .
the general feature of the relativistic calculations is that they provide the positive value of the difference @xmath447 in the matter with the neutron excess .
also the proton effective masses are above the neutron ones in this case , i.e. , @xmath448 .
our calculations show the same tendency . as to the quantitative results ,
our values of @xmath449 and @xmath450 appear to be about twice smaller than those , obtained in @xcite .
our value of the effective mass splitting is also about two times smaller than the result of @xcite but is only 30% smaller than that of @xcite .
however , we find somewhat smaller discrepancy with the relativistic brueckner hartree fock ( rbhf ) calculations , presented in @xcite .
they found @xmath451mev and @xmath452mev at @xmath453 ( fig . 3 of @xcite ) , while our results are @xmath454mev , @xmath455mev .
another rbhf analysis @xcite provided the results , which are very close to our ones .
one can extract the values @xmath456mev , and @xmath457mev at @xmath458 from figs .
10 and 11 of @xcite .
our values are 80 mev and 55 mev correspondingly .
note , however , that the split of the effective masses , obtained in @xcite is due to the exchange effects only .
the nonrelativistic calculations , carried out in various approaches @xcite provide @xmath459mev at @xmath370 and @xmath438 .
this is consistent with the earlier calculations @xcite .
our value is 40 mev at @xmath370 and @xmath438 .
another important parameter is the symmetry energy eq .
( [ 130 ] ) .
note that we calculate the quantity @xmath460 which is the true contribution to the energy per nucleon .
using eq .
( [ 112 ] ) we can express @xmath461 .
the value @xmath462 thus being the true contribution of the isovector forces .
our value is @xmath463 mev .
this is close to the one , obtained in @xcite .
the @xmath445 law is true up to @xmath464 with 10% accuracy in agreement with @xcite . to find the total contribution of the potential energy one must include the term caused by the exclusion effect , mentioned above
this adds 9 mev to @xmath443 @xcite . including also the contribution of the kinetic energy , we obtain @xmath465mev .
the various calculations of this parameter provide the values around 30 mev @xcite@xcite .
thus , our result agrees with those , obtained by nuclear physics methods . * v. summary * we expressed the vector and scalar self - energies of a nucleon in asymmetric nuclear matter as function of density @xmath66 and of the asymmetry parameter @xmath99 .
we presented the nucleon characteristics in terms of the in - medium expectation values of qcd operators .
the main ingredients are the nonlocal vector condensates @xmath466 , the scalar condensates @xmath467 and the four - quark condensates .
the local vector condensates are calculated easily .
the nonlocality of the vector condensates is expressed in terms of the nucleon structure functions .
the scalar condensate @xmath468 is presented in terms of the observable sigma term .
the scalar condensate @xmath469 and the four - quark condensates are calculated in framework of perturbative chiral quark model ( pcqm ) . although we treat the condensates in the gas approximation , the nucleon characteristics are not linear in density .
the corresponding equations ( [ 127a ] ) , ( [ 127 ] ) are analogous to the equations of nuclear physics beyond the mean field approximation . also eqs .
( [ 127a ] ) , ( [ 127 ] ) provide explicit expression of the nucleon characteristics in terms of the qcd condensates .
the successive inclusion of the ope terms in the lhs of the sum rules finds direct analogs in the meson - exchange description of the interactions of the nucleon in nuclear matter .
the lowest order ope terms correspond to the exchanges by the vector and ( effective ) scalar mesons with the point - like vertices of the interactions .
the higher order terms correspond to the nonlocal structure of nucleon - meson vertices , including the anomalous lorentz structures , and to the exchanges by the strongly correlated four - quark systems .
a possible interpretation of the latter contributions is a local two - meson exchange ( or the exchanges by the four - quark mesons , if there are any @xcite ) , figs . 1 , 2
. we obtained the functions @xmath329 and @xmath330 for all values of @xmath99 .
we calculated also the single - particle potential energy @xmath375 .
the results are presented in figs .
3 - 6 .
note , that we did not need phenomenological parameters of the nucleon meson interactions .
we used the condensates which have been either calculated , or expressed in terms of the observables .
for the four - quark condensates we used , however , the already known input parameters of pcqm . while including the condensates of the higher dimension in succession , we found the direct analogs of the meson
nucleon exchange mechanisms of nuclear physics .
our results for the nucleon self - energies are in reasonable agreement with the results of nuclear physics .
the value of the symmetry energy is close to the one obtained by the nuclear physics methods .
the work was supported by the grants rfbr - 03 - 02 - 17724 and rsgss - 1124.2003.2 .
* appendix a * here we discuss the isotopic structure of the nucleon current ( [ 28 ] ) .
the current ( [ 28 ] ) does not have a definite isospin .
this is not important for the calculations in a medium with the isospin i=0 , e.g. , in vacuum or in symmetric nuclear matter . in these cases
the expectation values of @xmath470 and @xmath471 operators have equal values ( the same is true for more complicated condensates ) .
this reasoning does not work for asymmetric matter .
therefore it is reasonable to start with the pure i=1/2 current @xmath472 @xmath473 \eqno{(a1)}\ ] ] however it turns out that even in asymmetric matter an explicit calculation of the polarization operator by using the current ( a1 ) provides the same result as that with the original ioffe current ( [ 28 ] ) .
this is due to the fermi statistics of the quarks . in the ope the current ( [ 28 ] )
annihilates a local three - quark state .
the coordinate function of the three quarks is symmetric .
the color structure is antisymmetric as well as the spin structure of the spin three quark system ( clearly the current ( [ 28 ] ) has spin 1/2 ) .
thus we need the antisymmetric isospin wave function . due to the fermi statistic of the fermions
any symmetric isospin component will vanish after the account for all possible permutations of the quarks .
this means that only the isospin 1/2 component actually survives in the ope expansion of the polarization operator ( [ 27 ] ) corresponding to the of current ( [ 28 ] ) . in other words
there is a relation between the contributions of the first and two last terms of the rhs of eq .
( a1 ) .
this provides a possibility to use ( 28 ) for the calculations in isospin asymmetric matter . * appendix b * the integrals in the rhs of eq .
( [ 84 ] ) can be evaluated by using the formula @xmath474\int\frac{d^4x}{x^6}e^{i(q'x ) } \eqno{(b1)}\ ] ] with @xmath475 .
the last factor is @xmath476 here the dots denote the terms which will be killed by the borel transform .
thus the lhs of the sum rules contains the contribution of the form @xmath477 with @xmath478 , @xmath479 .
we can present @xmath480 the first term on the rhs contains the standard logarithmic factor with the cut , running to infinity .
it is described by the `` pole + continuum '' model in a standard way .
the second term contains a finite cut .
it describes the singularities in the @xmath320-channel , caused by the nonlocal structure of the vector condensate , corresponding to the exchange terms on the rhs of the sum rules .
we neglect such contributions , thus coming to the hartree description of the nucleon in nuclear matter .
* appendix c * here we calculate the contribution of the four - quark condensates to @xmath11 structure of the polarization operator @xmath28 .
the contribution is provided by the scalar - vector condensate @xmath481 defined by eq .
( [ 90 ] ) with @xmath482 .
the first two terms determine the `` factorized '' contribution , which can be written as @xmath483 with the vector @xmath484 defined by eq .
( [ 64 ] ) . neglecting for a while the nonlocal contributions
, we put @xmath485 . in this approximation @xmath486 thus corresponding to eq .
( [ 100 ] ) with @xmath487 , @xmath488 .
the `` internal '' contributions , provided by the second terms in the rhs of eqs .
( [ 90 ] ) and ( [ 102 ] ) are also proportional to @xmath489 , with the coefficients @xmath490 determined in @xcite .
the factorized local terms and the internal terms provide the rhs of the last equality of eq .
( [ 100 ] ) .
taking into account the @xmath385-dependence of the vector @xmath484 in eq .
( c1 ) we obtain a more accurate expression for the `` factorized '' contribution , which contains the second moments of the structure functions and the terms of the @xmath205 expansion of the integrand @xmath491 @xmath492 .
\eqno{(c3)}\ ] ] the notations are explained in eqs .
( [ 79 ] ) and ( [ 81 ] ) , @xmath493 gev@xmath494 ( [ 41 ] ) , ( [ 42 ] ) . * appendix d * the functions @xmath400 and @xmath401 introduced by eq .
( [ 112 ] ) can be approximated by the polynomials of the second order : @xmath495 with @xmath496 .
the values of the coefficients @xmath497 are presented in table 1 .
this leads to parametrization of the proton potential energy @xmath498 \mbox{gev } .
\eqno{(d2)}\ ] ] * table * the values of the coefficients @xmath497 ( gev ) of the polynomials @xmath499 ( d1 ) , which approximate the nucleon self - energies . [ cols=">,>,>,>",options="header " , ] 99 m.a .
shifman , a.i .
vainshtein and v.i .
zakharov , nucl .
phys .
b * 147 * , 385 , 448 , 519 ( 1979 ) .
b.l .
ioffe , nucl .
phys .
b * 188 * , 317 ( 1981 ) .
b.l .
ioffe and a.v .
smilga , nucl .
phys .
b * 232 * , 109 ( 1984 ) .
m.a .
shifman , _ `` vacuum structure and qcd sum rules '' _ ( north holland , amsterdam , 1992 ) .
l.j .
reinders , h. rubinstein , s. yazaki , phys .
rep . * 127 * , 1 ( 1985 ) .
k.g .
wilson , phys . rev . * 179 * , 1499 ( 1969 ) . e.g. drukarev and e.m .
levin , jetp lett .
* 48 * , 338 ( 1988 ) ; sov .
phys .
jetp * 68 * , 680 ( 1989 ) .
e.g. drukarev and e.m .
levin , nucl .
phys .
a * 511 * , 679 ( 1990 ) ; prog .
part .
. phys .
* 27 * , 77 ( 1991 ) .
e.g. drukarev and m.g .
ryskin , nucl .
phys .
a * 578 * , 333 ( 1994 ) .
e.g. drukarev , m.g .
ryskin and v.a .
sadovnikova , prog .
part .
. phys . * 47 * , 73 ( 2001 ) .
j. gasser , ann .
phys . * 136 * , 62 ( 1981 ) .
u. weidner _
et al . _ ,
phys .
rev .
lett . * 58 * , 648 ( 1987 ) .
ulf - g .
meisner , g. smith , hep - ph/0011277 .
m.a .
shifman , a.i .
vainshtein and v.i .
zakharov , phys .
lett .
b * 178 * , 443 ( 1978 ) .
x. jin , t.d .
cohen , r.j .
furnstahl , and d.k .
griegel , phys .
rev .
c * 49 * , 464 ( 1994 ) .
e.g. drukarev , m.g .
ryskin , v.a .
sadovnikova , th .
gutsche and amand faessler , phys .
rev .
c ( in press ) . e.g. drukarev , m.g .
ryskin , v.a .
sadovnikova , v.e .
lyubovitskij , th .
gutsche and a. faessler , phys .
rev .
d * 68 * , 054021 ( 2003 ) .
t. gutsche and d. robson , phys .
lett .
b * 229 * , 333 ( 1989 ) .
v.e .
lyubovitskij , th .
gutsche and a. faessler , phys .
rev .
c * 64 * , 065203 ( 2001 ) .
v.e .
lyubovitskij , th .
gutsche , a. faessler and e.g. drukarev , phys .
rev .
d * 63 * , 054026 ( 2001 ) .
v.e .
lyubovitskij , th .
gutsche , a. faessler and r. vihn mau , phys .
rev .
c * 65 * , 025202 ( 2002 ) ; + v.e .
lyubovitskij , p. wang , th .
gutsche and a. faessler , phys .
rev .
c * 66 * , 055204 ( 2002 ) .
l.s .
celenza and c.m .
shakin , _ `` relativistic nuclear physics '' _ ( world scientific , philadelphia , 1986 ) . f.j .
yndurain , _
`` quantum chromodynamics '' _ , ( springer verlag .
ny , 1983 ) . c. itzykson and j .- b.zuber , _ `` quantum field theory '' _ , ( mcgraw hill , ny , 1980 ) .
v.m .
braun and a.v .
kolesnichenko , nucl .
phys .
b * 283 * , 723 ( 1987 ) .
x. jin , t.d .
cohen , r.j .
furnstahl , and d.k .
griegel , phys .
rev .
c * 47 * , 2882 ( 1993 ) .
e.g. drukarev , m.g .
ryskin , v.a .
sadovnikova and a. faessler , phys .
rev .
c * 65 * , 074051 ( 2002 ) ; e : * 66 * 039903 ( 2002 ) .
r.d .
carlitz and d.v .
creamer , ann .
phys . * 118 * , 429 ( 1979 ) ; + d.i .
diakonov and v.yu .
petrov , nucl .
phys .
b * 272 * , 457 ( 1986 ) .
j. gasser , h. leutwyler and m.e .
sainio , phys .
lett .
b * 253 * , 252 , 260 ( 1991 ) .
a.d .
martin , r.g .
roberts and w.j .
stirling , phys .
lett .
b * 387 * , 419 ( 1996 ) .
a.l .
fetter and j.d .
walecka , _
`` quantum theory of many - particle systems '' _ ( mcgraw - hill , new - york , 1971 ) .
j.v .
noble , nucl .
phys , a * 329 * , 354 ( 1979 )
. m. lopez - quelle , s. marcos , r. niembro , a. bouyssy and n.v .
giai , nucl .
phys .
a * 483 * , 479 ( 1988 ) .
b. liu , v. greco , v. baran , m. colonna and m. di toro , phys .
rev .
c * 65 * , 045201 ( 2002 ) .
v. greco , m. colonna , m. di toro , g. fabbri and f. matera , phys .
rev .
c * 64 * , 045203 ( 2001 ) .
f. de jong and h. lenske , phys .
rev .
c * 57 * , 3099 ( 1998 ) .
h. huber , f. weber and m.k .
weigel , phys .
rev .
c * 51 * , 1790 ( 1995 ) .
i. bombaci and u. lombardo , phys .
rev .
c * 44 * , 1892 ( 1991 ) .
w. zuo , i. bombaci , u. lombardo , phys .
rev .
c * 60 * , 024605 ( 1999 ) .
a. bohr and b. mottelson ,
_ `` nuclear structure '' _ , v.1 ( benjamin , ny , 1969 ) .
r. malfliet , prog .
part .
. phys .
* 21 * , 207 ( 1988 ) .
d. p. roy , j.phys .
g * 30 * , r113 ( 2004 ) . |
dms are semiconductors where a fraction @xmath30.01 - 0.1 ) of nonmagnetic elements is replaced by magnetic @xmath4 ions .
the class of iii - v dms has recently attracted considerable attention after the experimental observation of high @xmath0 s , due to significant improvements in molecular beam epitaxy techniques .
@xcite these compounds can play an important role in spintronic devices , and they represent a challenge to theory due to the combined presence of correlations and disorder . using a one - band model for dms , ( i ) the weak - coupling quadratic dependence of @xmath0 with @xmath2 is captured correctly by dmft @xcite , which also provides bulk limit results , while ( ii ) the mc techniques on finite clusters properly handle the random mn distribution and unveiled the reduction of @xmath5 at large @xmath2 due to carrier localization.@xcite for @xmath2 comparable to the hopping @xmath6 , both techniques reached similar conclusions , showing that they can complement towards a comprehensive understanding of dms .
however , much theoretical work remains to be done , particularly the consideration of the several bands active in real dms materials beyond mean field approximations.@xcite in this paper , we study a two - band model for dms at realistic low dopings @xmath7@xmath8@xmath9 , using a combination of @xmath10 techniques that gives unique characteristics to our work . to properly analyze the intermediate and strong coupling @xmath2 regime ,
mc simulations that correctly handle the random mn distribution are crucial . here , mc results for dms models including more than one band are presented for the first time.@xcite on the dmft side , pioneering calculations for two - band double - exchange hamiltonians already exist.@xcite however , those calculations focused on special cases , while our current dmft effort is more general , with two @xmath11=1/2 bands and arbitrary couplings , hoppings , and carrier densities . one of our main results is that @xmath0 can be substantially raised by considering multiband systems since at intermediate couplings the maximal @xmath0 at carrier filling @xmath1@xmath12@xmath7 is approximately twice larger than the highest @xmath5 obtained in the single - band model at filling @xmath1@xmath12@xmath13 ( note that the alternative notation @xmath14=@xmath15 , where @xmath16 is the hole density as a fraction of @xmath1 , will be used in some portions of the paper ) .
the qualitative reason is that the ib cooperate to raise @xmath0 for values of the chemical potential @xmath17 where those bands partially or fully overlap.@xcite the paper is organized as follows : section ii describes the model ; the dmft results are presented in section iii , while section iv is devoted to the monte carlo study .
conclusions and final remarks are in section v.
it is known that in @xmath4-doped @xmath18 , the @xmath4 ions substitute for @xmath19 cations and contribute itinerant holes to the valence band .
the @xmath4 ions have a half - filled @xmath20-shell which acts as a @xmath21 local moment .
due to a strong spin - orbit ( so ) interaction , the angular momentum @xmath22 of the @xmath1-like valence bands mixes with the hole spin degree of freedom @xmath23 and produces low- and high - energy bands with angular momentum @xmath24 and @xmath25 , respectively .
a robust so split between these bands causes the holes to populate the @xmath26 state , which itself is split by the crystal field into a @xmath27 band with heavy holes and a @xmath28 band with light holes .
this is the reason why we choose to study two bands since this is the relevant number of orbitals in most iii - v dms . since we do not work in a @xmath29 basis
our hamiltonian does not capture the orbital mixing in the hund term.@xcite however , we roughly consider the diagonal so effects in the magnetic interactions by allowing different values of @xmath2 in the two orbitals considered .
the simple two - band model for dms used here is given by the hamiltonian @xmath30 where @xmath31=1,2 is the band index ( not to be confused with angular momentum ) , @xmath32 label sites ( nearest neighbors for the hopping term ) , @xmath33 creates a hole at site @xmath34 in the band @xmath31 , @xmath35 is the spin- operator of the mobile hole ( @xmath36 = pauli vector ) , @xmath37 and @xmath38 are spin indices , @xmath39 is the coupling between the core spin and the electrons of band @xmath31 , and @xmath40 is the spin of the localized @xmath4 ion at randomly selected sites @xmath41 , assumed classical in the mc simulations .
@xmath42 is the hopping term in band @xmath31 .
the inter - band hopping @xmath43 ( @xmath44 ) is zero at the nearest neighbor level in cubic lattices @xcite .
even if @xmath43 is explicitly added , the conclusions are similar as reported here @xcite .
while real dms materials have zincblende ( zb ) structures , in this first nonperturbative study of a multiband dms model the simplicity of a cubic lattice allows us to focus on the dominant qualitative tendencies , a first step toward future quantitative studies with realistic zb lattices .
the model will be studied using dmft and mc techniques .
within dmft the self - energy is assumed to be local , @xmath45 , assumption valid exactly only in infinite dimensions .
the information about the hopping of carriers on and off lattice sites , which are magnetic ( with probability @xmath7 ) or nonmagnetic ( with probability @xmath9-@xmath7 ) , is in the bare green s function ( gf ) @xmath46 @xcite .
the full gf @xmath47 can be solved by integration with the result : @xmath48^{-1 } \rangle+(1-x)\langle{\mathcal{g}}_{0}(i\omega_{n})\rangle$ ] , where @xmath49=@xmath50 are matsubara frequencies .
the average @xmath51=@xmath52 is over the local moment @xmath53 orientations , with probability @xmath54 .
the equation above was derived for each band , and the equations set for the gf were solved on a bethe lattice with a bare semicircular noninteracting density of states @xmath55=@xmath56 . in this case ,
@xmath57 , where @xmath58=@xmath59 , and @xmath60=@xmath61 .
@xmath5 is obtained by linearizing the bare inverse gf with respect to the local order parameter @xmath62 .
to first order in @xmath63 , the implicit equation for @xmath0 is : @xmath64^{3 } + ( 48j_{l}^{2}-3w^{2}_{l})[b^{2}_{l}(i\omega_{n})-j_{l}^{2}]^{2}-5xj_{l}^{2}w_{l}^{2}[b^{2}_{l}(i\omega_{n})-j_{l}^{2}]-2xj_{l}^{4}w_{l}^{2}}=1,\ ] ] where @xmath65 satisfies a fourth - order - degree equation : @xmath66.@xcite our formula was tested in different cases : ( 1 ) for @xmath67=@xmath68 , we reproduce the one - band results of ref . , and ( 2 ) at @xmath7=1 , @xmath69=0 , and @xmath70 we reproduced the results of ref .
@xmath0 , contained in the matsubara frequencies , is obtained from eq.([tc ] ) numerically .
the equations for @xmath71 , rewritten in real frequency space via @xmath72 , give the interacting @xmath55 for each band at zero temperature when @xmath73.@xcite the solutions of these equations depend crucially on the ratio @xmath74 . the critical value for the formation of well - defined ib ,
corresponding to carrier spins locally parallel to @xmath4 spins , is @xmath75 .
below , the domain @xmath76 is referred to as ` weak coupling ' , while @xmath77 is the ` intermediate coupling ' .
the most interesting physics is observed at the boundary between these two regimes , i.e. when the ib are not completely separated from the valence bands . in fig .
[ fig1 ] , we show @xmath0 vs. @xmath1 , for different ratios @xmath78 and at fixed @xmath79=1 and @xmath80 , situation corresponding to the existence of a well - defined @xmath31=@xmath9 ib ( although @xmath1@xmath81@xmath7 in real dms , in this paper the case @xmath1@xmath82@xmath7 will also be studied for completeness , as done in ref . ) .
the inset shows the total interacting dos evolution .
the ib overlap if @xmath83 .
if the ib do not overlap , then each one determines @xmath5 separately , causing the double - peak structure observed for some @xmath84 ratios . the band with the largest @xmath74 is filled first , for smaller @xmath17 s .
at all @xmath1 s , we found that @xmath0 is maximum when @xmath85 , namely when the ib fully overlap .
the dependence of @xmath0 on @xmath78 at fixed @xmath1 is in fig .
[ fig2](a ) .
the peak value is achieved when the bands fully overlap ( i.e. at @xmath78=1 ) .
once the bands decouple , the value for @xmath0 matches one - band model results .
let us consider now how changes in bandwidths influence @xmath0 . in fig .
[ fig2](c ) , we show @xmath0 vs. @xmath1 parametric with @xmath86 , at fixed @xmath80 ( intermediate coupling ) , and with @xmath85.@xcite at small @xmath86 the second ib shall be located in a region of @xmath87 smaller ( i.e. farther from the valence bands ) , than the energy interval occupied by the @xmath88 ib .
hence , the @xmath31=2 ib will be the first to be filled .
decreasing @xmath89 , the second band moves to the right on the @xmath87 axis , towards the location of the first band . while the bands are still separated , each gives its own contribution to @xmath0 .
the curves with @xmath90 and @xmath91 correspond to decoupled ib , while those with @xmath92 and @xmath93 correspond to partially overlapping bands .
again , @xmath0 is maximized at all fillings when the bands fully overlap ( @xmath86=@xmath9 ) , in good agreement with ref .. although @xmath86=@xmath9 is not realistic in dms , mn doping materials with a relatively small heavy - light mass ratio will favor a higher @xmath5 .
once established that @xmath0 is maximal for all @xmath1 when @xmath78=@xmath9 and @xmath86=@xmath9 , let us analyze @xmath0 vs. @xmath1 when @xmath94 varies .
the results for @xmath0 , and total interacting dos , are in fig.[fig3 ] . at small coupling @xmath95 , @xmath0 is small , flat , and much extended on the @xmath1 axis , qualitatively similar to the one - band results.@xcite however , at intermediate coupling , @xmath0 is nonzero in the range from @xmath96 to @xmath97 , adopting a parabolic form with the maximum at @xmath98 , in contrast with the one - band model which gives a null @xmath0 when @xmath98 .
the explanation is straightforward : at @xmath99 in the one - band model the ib is fully occupied leading to a vanishing @xmath0 , but for the same @xmath1 in the two - band model both bands are half filled , which ultimately leads to the highest value for @xmath0 .
the @xmath0 dependence on the ratio @xmath94 at some fillings @xmath1 is displayed in fig.[fig2](b ) . at small coupling
, @xmath0 correctly increases quadratically with @xmath94 , but at strong coupling @xmath0 incorrectly continues growing , result which will be improved upon by the mc simulations shown below .
the hamiltonian eq . ( [ ham ] ) was also studied numerically using mc techniques similar to those applied to mn - oxide investigations.@xcite the fermionic sector is treated exactly , while a mc simulation is applied to the classical localized spins .
cubic lattices with @xmath100 and @xmath101 sites were investigated .
these lattice sizes have been shown to be sufficient for the comparison with dmft results and to unveil general trends .
in addition , the figures show only small variations for the @xmath0 estimations using the two lattices .
however , if more sophisticated quantitative analysis is required , clearly larger systems will be needed .
one may suspect that actually the number of mn spins may regulate the size effects , rather than the number of sites .
for the small values of @xmath7 used in our study , the number of mn spins is also very small and serious size effects could be expected . however , in practice this does not seem to occur , and moreover the comparison with dmft shows similar results using both techniques . perhaps in the small j s regime , the delocalized nature of the carriers smears the effects caused by the actual location of the mn spins .
these issues deserve further study , but for our purposes of unveiling trends in the multi - parameter space of dms materials , the lattices here used are sufficient . returning to the numerical data , the spin magnetization is the order parameter that was used to detect the ferromagnetic transition.@xcite all quantities are in units of @xmath102=1 , and the density of magnetic impurities is @xmath103 . in fig .
4(a ) , typical magnetization curves are presented at several carrier densities @xmath14 , and for @xmath104=@xmath69=@xmath2=4 and @xmath105=@xmath9 . in excellent agreement with dmft
, it was observed that the estimated @xmath5 is the highest for @xmath14=@xmath9 .
the value of @xmath2 used maximizes the critical temperature , and it was confirmed that it corresponds to the case where the ib are about to become separated from the valence band . in fig .
4(b ) , it is shown how @xmath5 increases with @xmath2 up to @xmath2=@xmath106 , in agreement with dmft .
the strong coupling behavior is nevertheless different since @xmath5 _ decreases _ at large @xmath2 .
this is caused by hole localization in strong coupling , @xcite beyond the capabilities of dmft .
the dependence of @xmath5 on the ratio of band hoppings is in fig .
4(c ) , varying @xmath14 .
these results are again in good qualitative agreement with dmft ( fig .
the maximum @xmath5 for all values of @xmath14 occurs when @xmath107=@xmath9 . however , when @xmath105 is very different from @xmath102 the development of magnetism is regulated by only one of the ib and the results are similar to those obtained with a single band model , as shown in the curves for @xmath105=@xmath108 and @xmath109 in fig
. 4(c ) . for @xmath107
closer to 1 , a partial overlap of the ib occurs and a hump in @xmath5 develops at @xmath14=@xmath9 ( see curve for @xmath105=@xmath93 ) . in fig .
4(d ) we show that , once @xmath107 is optimized , a similar finite @xmath2 maximizes @xmath5 for several @xmath14 s . in all cases , the optimal @xmath2 best balances the weak coupling behavior , with mobile holes not much affected by the interaction with the mn ions , and the strong coupling region where the hole spins strongly align with the mn spins , becoming localized . this `` sweet spot ''
is achieved when the ib are about to be separated from the valence bands .
@xmath5 vs. @xmath14 , at several ratios @xmath84 and for @xmath102=@xmath105 is presented in fig .
5(a ) . in agreement with dmft ( fig . 1 )
, @xmath5 is maximized for all values of @xmath14 if @xmath69=@xmath104 , with the highest value at @xmath14=@xmath9 .
overall , there is an excellent agreement with dmft , as described in the caption.@xcite @xmath5 vs. @xmath14 for several @xmath104=@xmath69=@xmath2 is in fig .
5(b ) . at small @xmath2 ,
again the mc results resemble those obtained with dmft ( fig . 3 ) . for , e.g. ,
@xmath2=@xmath110 the ib are not formed yet ( inset ) . in this regime , @xmath5 remains finite , although small , even for @xmath14 larger than 2 .
increasing @xmath2 , a nonzero @xmath5 is obtained only for @xmath14 between 0 and 2 , due to ib formation .
@xmath5 reaches a maximum at @xmath2=4 .
we have carried out the first study of a multiband model for dms using a powerful combination of nonperturbative techniques , dmft and mc .
we found the parameter regime that maximizes @xmath5 .
this happens at intermediate couplings and for all hole densities when @xmath111=@xmath9 and @xmath112=@xmath9 .
the maximum @xmath0 is obtained at @xmath98 , in contrast with the one - band model which has a vanishing @xmath5 at the same doping .
in addition , @xmath0 at filling @xmath113 in the one - band case is smaller than with two bands by a factor @xmath1142 . in view of the simplicity of the main results ,
it is clear that adding an extra band to the calculations ( which is relevant for system with negligible so , but considerably raises the cpu cost ) will only lead to a further increase in @xmath0 when all the ib overlap .
the excellent agreement dmft - mc is somewhat surprising due to the fact that monte carlo considers the influence of the random location of the mn sites much better than dmft .
however , at small and intermediate j s , the carriers can be sufficiently delocalized that a smearing effect may occurs and considering the quenched disorder only in average appears to be sufficient .
certainly at large j s the mc and dmft methods give totally different answers , with mc capturing the correct localization result .
the approach described here is also quantitative .
in fact , using gaas realistic parameters such as @xmath1=@xmath115 ( @xmath16=@xmath116/mn ) , a bandwidth @xmath114 10 ev ( @xmath102@xmath117 2.5 ev ) , @xmath105=(1/9)@xmath102 , and assuming @xmath118=@xmath9 and @xmath104=@xmath69 , we obtain @xmath119 , i.e. within the experimental range .
while this excellent agreement with experiment@xcite may be accidental , the trends are reliable and the result improves upon single - band estimations .
moreover , for optimal @xmath107=1 and @xmath120=2 , the @xmath5 raises to @xmath121 , even at small @xmath14=0.1 , setting the upper bound for dms under a two - band model description using a cubic lattice . the general qualitative picture presented here can be used to search for dms with even higher @xmath5 s than currently known .
our results suggest that semiconductors with the smallest heavy to light hole mass ratio , such as alas , could have the highest @xmath5 if the couplings @xmath2 could be tuned to its optimal value .
the present effort paves the way toward future nonperturbative studies of dms models using realistic zb lattices , and points toward procedures to further increase the curie temperatures .
we acknowledge conversations with k. aryanpour , m. berciu , r.s .
fishman , j. moreno , and j. schliemann .
this research is supported by grant nsf - dmr-0443144 .
ornl is managed by ut - battelle , llc under contract no .
de - ac05 - 00or22725 .
mc simulations used the spf program ( http://mri-fre.ornl.gov/spf ) .
h. ohno , science * 281 * , 951 ( 1998 ) ; t. dietl , h. ohno , f. matsukura , j. cibert , and d. ferrand , science , * 287 * , 1019 ( 2000 ) ; s. j. potashnik , k.c .
chun , j.j .
berry , n. samarth , and p. schiffer , appl .
. lett . * 79 * , 1495 ( 2001 ) .
a. chattopadhyay , s. das sarma , and a.j .
millis , phys .
lett . * 87 * , 227202 ( 2001 ) , and refs . therein . for a dmft review ,
see a. georges , g. kotliar , w. krauth , and m.j .
rozenberg , rev .
mod . phys . * 68 * , 13 ( 1996 ) .
j. sinova , t. jungwirth , s .-
eric yang , j. kucera , and a. h. macdonald , phys .
b*66 * , 041202 ( 2002 ) ; t. jungwirth , j. konig , j. sinova , j. kucera , and a.h .
macdonald , _ ibid _ * 66 * , 012402 ( 2002 ) . in j. schliemann , j. konig and a.h .
macdonald , phys .
rev . b*64 * , 165201 ( 2001 ) , mc studies of multiband models at small @xmath2 were presented , but without using an underlying lattice , and replacing the fermion trace by a ground state expectation value .
the term `` impurity band '' is widely used in the dms literature ( see e.g. m. berciu and r. n. bhatt , phys .
lett . * 87 * , 107203 ( 2001 ) ; j. okabayashi , a. kimura , o. rader , t. mizokawa , a. fujimori , t. hayashi , and m. tanaka , phys .
b*64 * , 125304 ( 2001 ) ) , although it should not be confused with the same term used in nonmagnetic doped semiconductor literature .
perhaps `` exchange split valence band '' could be a better terminology .
g. zarnd and b. jank , phys .
lett . * 89 * , 047201 ( 2002 ) . in ref .
it was shown that the frustration can be turned off by giving equal masses to the light and heavy sectors .
our emphasis has been on this particular case , since under this circumstance the @xmath0 is maximized .
the relative position of the bands is fixed by shifting numerically the @xmath31=2 valence band such that both valence bands start at the same energy , to mimic the degeneracy of the light and heavy bands of gaas at @xmath124 . | using dynamical mean field theory ( dmft ) and monte carlo ( mc ) simulations , we study the ferromagnetic transition temperature ( @xmath0 ) of a two - band model for diluted magnetic semiconductors ( dms ) , varying coupling constants , hopping parameters , and carrier densities . we found that @xmath0 is optimized at all fillings @xmath1 when both impurity bands ( ib ) fully overlap in the same energy range , namely when the exchange couplings @xmath2 and bandwidths are identical .
the optimal @xmath0 is found to be about twice larger than the maximum value obtained in the one - band model , showing the importance of multiband descriptions of dms at intermediate @xmath2 s . |
let @xmath2 be a homogeneous ideal in a polynomial ring @xmath3 $ ] over a field @xmath4 .
assume that @xmath2 is gorenstein , i.e. the quotient ring @xmath5 is gorenstein .
the general question whether @xmath2 has a ( possibly square free ) gorenstein initial ideal has been discussed recently by several authors for classical families of ideals , see @xcite . in particular , a positive answer is given for important classes of classical ideals :
ideals of minors @xcite , ideals of pfaffians @xcite and plcker relations @xcite . in these examples
the associated gorenstein initial ideals are actually square - free and the corresponding simpicial complexes have a beautiful combinatorial descriptions .
indeed , they provide a link to the theory of ( multi)-associahedra and to generalized cluster complexes ( see @xcite ) .
let @xmath6 $ ] be the polynomial ring in the variables @xmath7 over a field @xmath4 .
for @xmath8 we set @xmath9 and consider the generic @xmath10 symmetric matrix @xmath11 .
for @xmath12 we denote by @xmath13 the ideal generated by the @xmath14-minors of @xmath15 .
it is known that @xmath16 is a cohen - macaulay normal domain and that it is gorenstein if and only if @xmath17 is even , see @xcite .
it is proved in @xcite and in @xcite that the @xmath14-minors of @xmath15 are a grbner bases with respect to the lexicographic order induced by @xmath18 the corresponding initial ideal is square - free and cohen - macaulay and it is even the stanley - reisner ideal of a shellable simplicial complex .
but , apart from few exceptions ( @xmath19 ) , these initial ideals are not gorenstein because they do not have the right number of cone points " , see @xcite for details .
on the other hand , a gorenstein square free initial ideal of @xmath20 has been given in @xcite and ( implicitly ) in @xcite .
the goal of this paper is to treat the case of the ideal @xmath21 of minors of size @xmath22 of a symmetric matrix of variables of size @xmath10 . in the remaining cases , i.e. @xmath23 and @xmath17
even , we have not been able to identify a gorenstein initial ideal for @xmath13 .
indeed , we do not even have a guess . returning to the case @xmath24
, we will actually prove that the minors of size @xmath22 of @xmath15 form a grbner basis of the ideal @xmath21 with respect to a suitable reverse lexicographic order @xmath25 and that the corresponding initial ideal @xmath26 is square - free and gorenstein .
furthermore , we will show in section [ r - covered ] that the simplicial complex associated to @xmath26 is the boundary complex of a cyclic polytope .
typically the betti numbers of initial ideals are bigger than the betti numbers of the original ideal , and this happens also for determinantal ideals whose initial ideal is gorenstein , see for instance the examples in @xcite .
this behaviour can be explained theoretically , at least for minors of order @xmath27 , using the logarithmic bounds for the regularity of a quadratic monomial ideals with first linear syzygies established in @xcite . in our case however , it turns out that the betti numbers of @xmath21 and @xmath26 actually coincide .
the reason is that the betti numbers of a compressed graded gorenstein @xmath28-algebra of even regularity just depend on the regularity and the codimension , see [ compressed ] . by observing that @xmath29 and @xmath30
are gorenstein compressed of regularity @xmath31 and codimension @xmath32 one concludes that @xmath33 . to sum up
, the goal of this note is to prove the following [ main ] let @xmath15 be the generic @xmath10 symmetric matrix .
then there exists a reverse lexicographic term order @xmath25 such that : 1 .
the @xmath0-minors of @xmath15 form a grbner basis of @xmath21 ; 2 .
@xmath34 is square - free and defines a gorenstein ring ; 3 .
@xmath35 for all @xmath36 ; 4 .
@xmath34 is an iterated cone over the stanley - reisner ideal of the boundary complex of a @xmath37-dimensional cyclic polytope on @xmath38 vertices .
in the following theorem we collect important results about the determinantal ring @xmath16 of a generic symmetric matrix @xmath15 of size @xmath10 proved by kutz @xcite , goto @xcite , harris and tu @xcite and conca @xcite .
[ pr : basic ] the ring @xmath16 is a cohen - macaulay normal domain .
it is gorenstein if and only if @xmath17 is even .
krull dimension , multiplicity , castelnuovo - mumford regularity and @xmath39-invariant of @xmath16 are given by the following formulas : @xmath40 in this paper we consider the case of the ideals of minors of size @xmath24 .
as a special case of [ pr : basic ] we have : [ co : n-2 ] the ring @xmath29 is a gorenstein normal domain .
its dimension , multiplicity , regularity and @xmath39-invariant are : @xmath41 [ compressed ] let @xmath42 be a gorenstein graded @xmath28-algebra of even castelnuovo - mumford regularity @xmath43 and codimension @xmath44
. then the @xmath45-polynomial @xmath46 satisfies the inequality @xmath47 for @xmath48 .
since @xmath49 , this gives an upper bound for the multiplicity of @xmath42 only in terms of @xmath44 and @xmath50 : @xmath51 and @xmath42 is said to be compressed if the equality holds .
in other words , @xmath42 is compressed if its h - polynomial is given by @xmath52 a simple computation shows that the minimal free resolution of a compressed gorenstein @xmath28-algebra of even regularity is pure ( i.e. only one shift in each homological position ) and hence its betti numbers just depend on @xmath50 and @xmath44
. explicit expressions for the betti numbers can be worked out , they can be found for example in @xcite .
we see from [ co : n-2 ] that @xmath29 is compressed of even regularity .
we obtain the following expressions for the @xmath45-polynomial and the betti numbers of @xmath29 .
[ betti ] the codimension of @xmath29 is @xmath32 , its @xmath45-polynomial is @xmath53 and its non - zero betti numbers are : @xmath54 and @xmath55
we define a term order on the monomials in @xmath56 : set @xmath57 consider a termorder @xmath25 given by the reverse lexicographic order on @xmath58 with the variables ordered as follows : @xmath59 in the sequel , for @xmath50-elements subsets @xmath60 and @xmath61 of @xmath62 we write @xmath63 $ ] for the minor of the generic symmetric matrix @xmath15 defined by selecting rows @xmath60 and columns @xmath61 in the given order .
note that we sometimes speak of the row or column of a minor by which we mean the row or column of the submatrix selected to compute the minor . for classifying the leading monomials of the @xmath0-minors
we first need some preparatory lemmas .
[ lem : small ] let @xmath64 and @xmath65 and @xmath66 be two sequences of distinct indices such that @xmath67 for @xmath68 . then @xmath69 = [ \a_{1}|\b_{1 } ] \cdots [ \a_s|\b_s].\ ] ] we prove the assertion by induction on @xmath50 .
the case @xmath70 is trivial and we can assume @xmath71 . set @xmath72 $ ] .
we expand @xmath73 along the @xmath50^th^ column : @xmath74[\a_1,\ldots , \a_{i-1 } , \a_{i+1 } , \ldots , \a_s|\b_1,\ldots , \b_{s-1}].\ ] ] by induction we have that @xmath75)=x_{\a_1\b_1 } \cdots x_{\a_{s-1 } \b_{s-1}}.\ ] ] by @xmath76 it follows that @xmath77 for @xmath78 and equality can only hold if @xmath79 and @xmath80 and @xmath81 . but this contradicts @xmath66 and hence @xmath82 for @xmath78 .
this implies that @xmath83\not\in d\cup u$ ] for every @xmath84 . since @xmath85 \in u$ ] and @xmath86)$ ] is a monomial in @xmath87 it follows from the choice of @xmath25 that @xmath88 takes the desired form .
[ lem : newlittle ] let @xmath89 .
assume we have two sequences @xmath90 of distinct indices
such that : * @xmath67 for @xmath91 , * @xmath92 and @xmath93 for @xmath94 and @xmath95 , * @xmath96 for @xmath97 , * @xmath98 $ ] is a monomial in @xmath87
. then @xmath99 & = & \\ { \mathrm{in}}_\prec [ a_1,\ldots , a_r|b_1,\ldots , b_r ] [ \a_{r+1}|\b_{r+1 } ] \cdots [ \a_s|\b_s ] . & & \end{aligned}\ ] ] we proceed by induction on @xmath100 . if @xmath101 then the assertion is trivially true .
assume @xmath102 .
set @xmath103 $ ] .
we expand @xmath73 along the @xmath50^th^ column : @xmath104[a_1,\ldots , a_{i-1 } , a_{i+1 } , \ldots , a_r , \a_{r+1},\ldots , \a_s|b_1,\ldots , b_r , \b_{r+1},\ldots , \b_{s-1}]+ \\ & & \sum_{i = r+1}^{s } \pm[\a_i \,|\,\b_s][a_1,\ldots , a_r , \a_{r+1 } , \ldots,\a_{i-1},\a_{i+1 } , \ldots , \a_s|b_1,\ldots , b_r , \b_{r+1},\ldots , \b_{s-1 } ] .
\end{aligned}\ ] ] for the summation index @xmath105 by induction we have @xmath106[a_1,\ldots , a_r , \a_{r+1 } , \ldots,\a_{s-1}|b_1,\ldots , b_r , \b_{r+1},\ldots , \b_{s-1 } ] ) & = & \\ { \mathrm{in}}_\prec ( [ \a_s|\b_s ] [ a_1,\ldots , a_r|b_1,\ldots , b_r ] [ \a_{r+1}|\b_{r+1 } ] \cdots [ \a_{s-1}|\b_{s-1 } ] ) . & & \end{aligned}\ ] ] which by our assumption on @xmath107)$ ] is a monomial in @xmath87 . note that @xmath108 $ ] do not lie in @xmath109 for @xmath94 , and @xmath110 $ ] also do not lie in @xmath109 for @xmath111 .
thus by the choice of @xmath25 it follows that @xmath112[a_1,\ldots , a_r , \a_{r+1 } , \ldots,\a_{s-1}|b_1,\ldots , b_r , \b_{r+1},\ldots , \b_{s-1 } ] \\ & = & { \mathrm{in}}_\prec ( [ a_1,\ldots , a_r|b_1,\ldots , b_r ] ) [ \a_{r+1}|\b_{r+1 } ] \cdots [ \a_s|\b_s ] .
\end{aligned}\ ] ] the same proof but expanding along the first row yields : [ lem : newlittle2 ]
let @xmath113 .
assume we have two sequences @xmath114 of distinct indices such that : * @xmath67 for @xmath115 , * @xmath116 and @xmath117 for @xmath118 and @xmath119 , * @xmath96 for @xmath120 , * @xmath121 $ ] is a monomial in @xmath87
. then @xmath122 & = & \\ \,[\a_{1}|\b_{1 } ] \cdots [ \a_{s - q}|\b_{s - q } ] { \mathrm{in}}_\prec [ a_{s - q+1},\ldots , a_s|b_{s - q+1},\ldots , b_s ] . & & \end{aligned}\ ] ] [ lem : special ] let @xmath123 , @xmath124 , @xmath125 for @xmath126 and @xmath127 , @xmath128 , @xmath129 for @xmath126 .
then for @xmath130 one has @xmath131 & = & x_{1,2 } x_{1,3 } x_{2,4 } \cdots x_{2\ell,2\ell+2}. \end{aligned}\ ] ] and for @xmath132 one has @xmath133 & = & x_{1,2 } x_{1,3 } x_{2,4 } \cdots x_{2\ell-1,2\ell+1}. \end{aligned}\ ] ] we proceed by induction on @xmath134 . for @xmath135
a direct computation yields the result .
assume @xmath136 .
first , we consider @xmath137 $ ] for @xmath132 .
the entries of the @xmath138^th^ row of @xmath137 $ ] are @xmath139 $ ] , @xmath140 $ ] for @xmath126 and @xmath141 $ ] for @xmath142 . since @xmath143 and @xmath144 we have
that @xmath139 $ ] , @xmath141 \not\in d \cup u$ ] for @xmath142 . for @xmath142 also @xmath140 \not \in d \cup u$ ] . thus expanding @xmath137
$ ] along its last row we obtain for all columns except for the @xmath138^th^ column a factor that does not lie @xmath109 . for the last column
we get the term @xmath145 [ a_1,\ldots , a_{2(\ell-1)+1}|b_1,\ldots , b_{2(\ell-1)+1}].\ ] ] by the induction hypothesis its initial term is a monomial in @xmath87 and thus it is the initial term @xmath137 $ ] .
this implies for @xmath134 .
analogously , consider the @xmath146^st^ column of @xmath147 $ ] .
its entries are @xmath148 $ ] , @xmath149 $ ] , @xmath150 and @xmath151 $ ] , @xmath152 .
since @xmath143 and @xmath144 we have that @xmath148 $ ] , @xmath149 \not \in d \cup u$ ] for @xmath152 . for @xmath153 also @xmath151 \not\in d \cup u$ ] .
thus expanding @xmath147 $ ] along its last column we obtain for all row except of the @xmath146^st^ row a factor that does not lie @xmath109 .
for the last row we get the term @xmath154 [ a_1,\ldots , a_{2\ell}|b_1,\ldots , b_{2\ell}].\ ] ] by for @xmath134 its initial term is a monomial in @xmath87 and therefore is the initial term of @xmath147 $ ] .
this then implies for @xmath134 .
[ lem : special2 ] let @xmath155 , @xmath156 , @xmath157 for @xmath126 and @xmath158 , @xmath159 , @xmath160 for @xmath126 .
then for @xmath132 @xmath161 & = & x_{n-2\ell-1,n-2\ell+1 } \cdots x_{n-3,n-1 } x_{n-2,n } x_{n-1,n } \end{aligned}\ ] ] and for @xmath130 @xmath162 & = & x_{n-2\ell , n-2\ell+1 } \cdots x_{n-3,n-1 } x_{n-2,n } x_{n-1,n}. \end{aligned}\ ] ] [ le : initial1 ] for a number @xmath64 let @xmath163 , @xmath164 be indices such that @xmath165 is a squarefree monomial with @xmath166 for every @xmath167 . then @xmath168 is the leading term of an @xmath50-minor of @xmath15 with respect to @xmath25 .
we distinguish cases according to the order relations among the @xmath169 and among the @xmath170 .
case 1 : @xmath65 and @xmath66 .
here the assertion follows from [ lem : small ] .
case 2 : @xmath171 with at least one equality and @xmath66 . in this situation
it follows that @xmath172 and @xmath173 , @xmath174 .
if @xmath175 for @xmath176 then by @xmath177 we must have @xmath178 and the assertion follows from [ lem : special ] .
thus we can assume that there is a @xmath179 such that @xmath180 .
we choose @xmath181 minimal with this property .
set @xmath182 . by [ lem : special ] there
exist @xmath183 , @xmath184 such that @xmath185 = m'$ ] . by the [ lem : special ] after possibly exchanging rows and columns we can assume @xmath186 for @xmath187
note that exchanging rows and columns does not change the minor since @xmath15 is symmetric . then by the choice of @xmath181 we have @xmath188 , thus @xmath189 and @xmath190\not\in d\cup u$ ] for @xmath187 , @xmath191 .
moreover , by @xmath192 for @xmath97 it follows that @xmath193 for @xmath97 .
therefore , we can apply [ lem : newlittle ] to @xmath194 and @xmath195 .
this shows that @xmath196 & = & \\ { \mathrm{in}}_\prec [ a_1,\ldots , a_r|b_1,\ldots , b_r ] [ \a_{r+1}|\b_{r+1 } ] \cdots [ \a_s|\b_s ] & = & \\ m ' x_{\a_{r+1 } \b_{r+1 } } \cdots x_{\a_s\b_s } & = & \\ x_{\a_1\b_1}x_{\a_2\b_2}\cdots x_{\a_r\b_r}x_{\a_{r+1 } \b_{r+1 } } \cdots x_{\a_s\b_s } & = & m. \end{aligned}\ ] ] case 3 : @xmath65 and @xmath197 with at least one equality . in this situation
it follows that @xmath198 and @xmath199 , @xmath200 .
if @xmath201 for @xmath202 then by @xmath177 we must have @xmath203 for @xmath204 and the assertion follows from [ lem : special2 ] .
thus we can assume that there is a @xmath205 such that @xmath206 .
we choose @xmath207 minimal with this property .
set @xmath208 . by [ lem : special2 ] there exist @xmath209 , @xmath210 such that @xmath211 = m'$ ] . by the [ lem : special ] after possibly exchanging rows and columns we can assume @xmath212 for @xmath213 . then by the choice of @xmath207 we have @xmath214 , thus @xmath215 and @xmath216\not\in d\cup u$ ] for @xmath213 and @xmath217 .
moreover , by @xmath192 for @xmath120 it follows that @xmath193 for @xmath120 .
therefore , we can apply [ lem : newlittle2 ] to @xmath218 and @xmath219 .
this shows that @xmath220 & = & \\
\,[\a_{1}|\b_{1 } ] \cdots [ \a_{s - q}|\b_{s - q } ] { \mathrm{in}}_\prec [ a_{s - q+1},\ldots , a_s|b_{s - q+1},\ldots , b_s ] & = & \\
x_{\a_{1 } \b_{1 } } \cdots x_{\a_{s - q}\b_{s - q } } m ' & = & \\
x_{\a_1,\b_1 } \cdots x_{\a_{s - q},\b_{s - q}}x_{\a_{s - q+1}\b_{s - q+1}}x_{\a_{s - q+2}\b_{s - q+2}}\cdots x_{\a_s\b_s } & = & m .
\end{aligned}\ ] ] case 4 : @xmath171 and @xmath197 with at least one equality in both .
in this situation it follows that @xmath172 and @xmath173 , @xmath174 and @xmath198 and @xmath199 , @xmath200 .
thus we can assume that there is a @xmath179 such that @xmath221 and there is a @xmath205 such that @xmath222 .
we choose @xmath181 and @xmath207 minimal with this property .
set @xmath182 . by [ lem : special ] there
exist @xmath183 , @xmath184 such that @xmath185 = m'$ ] .
after possibly exchanging rows and columns we can assume @xmath186 for @xmath187 .
moreover set @xmath223 . by [ lem : special2 ]
there exist @xmath209 , @xmath210 such that @xmath224 = m'',\ ] ] and after possibly exchanging rows and columns we can assume @xmath212 for @xmath213 . set @xmath225.\ ] ] note that the row and the column indices of @xmath73 are distinct since @xmath226 , @xmath227 for @xmath228 and @xmath229 , @xmath230 for @xmath231 .
we prove that @xmath232 , by induction on @xmath233 .
if @xmath234 , then @xmath235 $ ] . by expanding along the first row one
obtains : @xmath236m_1- [ 2 \,|\,3]m_2 + \sum_{j=3}^{s - q } \pm[2
\,|\,\b_j]m_j + \sum_{j = s - q+1}^{s } \pm[2 \,|\,b_j]n_j .\ ] ] note , that @xmath237\not\in d\cup u$ ] and since @xmath238 and @xmath239 , then @xmath240 $ ] and @xmath241 $ ] do not lie in @xmath242 for every @xmath243 . thus @xmath244m_1)=[2,1]{\mathrm{in}}(m_1)$ ] .
indeed , the only element from @xmath109 in the first row of @xmath245 is i@xmath246 $ ] .
hence , by expanding @xmath245 along the first row and by using case 3 one obtains : @xmath247{\mathrm{in}}([\a_{3},\ldots , \a_{s - q},a_{s - q+1},\ldots , a_s|\b_{3},\ldots , \b_{s - q},b_{s - q+1},\ldots , b_s])=m.\ ] ] now assume @xmath248 .
we consider the case when @xmath181 is even .
the case when @xmath181 odd can be treated in the same way .
one has : @xmath249\ ] ] with @xmath250 .
note that @xmath251 , @xmath252 for every @xmath243 and @xmath253 for all @xmath36 . by expanding @xmath73 along the @xmath254^th^ row one
obtains : @xmath255m_1- [ a_r \,|\,3]m_2 + \sum_{j=3}^{r } \pm[a_r \,|\,b_j]p_j+\sum_{j = r+1}^{s - q}\pm[a_r \,|\,\b_j]m_j+\sum_{j = s - q+1}^{s}\pm[a_r \,|\,b_j]n_j\ ] ] consider @xmath256 $ ] . by the assumption on the indices involved in @xmath73 ,
the only element of the @xmath257^st^ row in @xmath242 is @xmath258=[r , r-2]$ ] .
thus by expanding @xmath259 along the @xmath257^st^ row we get that @xmath260p_r)$ ] equals @xmath261[a_{r-1},b_{r-1}][2,1 , a_{3},\ldots , a_{r-2 } , \a_{r+1},\ldots , a_s|1,3,b_{3},\ldots , b_{r-2},\b_{r+1},\ldots,\b_{s - q},b_{s - q+1},\ldots , b_s]\ ] ] by the induction hypotheses .
now @xmath262\not\in d\cup u$ ] for @xmath263 and @xmath264\not\in d\cup u$ ] for @xmath231
. moreover , if @xmath265 the only @xmath264 $ ] that can be in @xmath242 are from the set @xmath266=[r-1,r+1],\ [ a_r , b_{r-2}]=[r-1,r-1],\ [ a_r , b_{r-4}]=[r-1,r-3]\}$ ] .
it follows that @xmath267m_1 + [ a_r \,|\,3]m_2+[a_r \,|\,b_{r-2}]p_{r-2}+[a_r \,|\,b_{r-4}]p_{r-4})\\ \ & = { \mathrm{in}}(m+[a_r \,|\,1]m_1 + [ a_r \,|\,3]m_2+[a_r \,|\,a_{r}]p_{r-2}+[a_r \,|\,a_r-2]p_{r-4 } ) .
\end{array}\ ] ] note , that since @xmath268 even implies @xmath269 . if @xmath270 , then @xmath271=[3,1]$ ] and @xmath272=[3,3]=[a_r , b_{r-2}]$ ] .
if @xmath273 , then @xmath274=[5,1]\not\in d\cup u$ ] and @xmath272=[a_r , b_{r-4}]$ ] .
if @xmath275 , then @xmath276 and @xmath274 $ ] and @xmath272 $ ] do not lie in @xmath242 .
therefore , @xmath277p_{r-2}+[a_r \,|\,a_r-2]p_{r-4 } ) & \mbox { if } r\ge 6\\ { \mathrm{in}}(m+ [ 3 \,|\ , 1]m_1+[3\,|\ , 3]p_{r-2 } ) & \mbox { if } r=4 \\ \end{array}. \right.\ ] ] in the following we prove that @xmath278 and @xmath279 involve at least one indeterminate not in @xmath242 , thus @xmath280{\mathrm{in}}(p_{r-2 } ) , [ a_r \,|\,a_r-2]{\mathrm{in}}(p_{r-4})$ ] and @xmath281{\mathrm{in}}(m_1)$ ] are larger than @xmath168 . from this the assertion follows . we will only treat the case of @xmath282
, an analogous reasoning will covers the cases @xmath283 and @xmath245 . for @xmath284
$ ] the only element of the @xmath257^st^ row which is in @xmath242 is @xmath258=[r , r-2]$ ] . expanding along the @xmath257^st^ row , we get @xmath285q$ ] , with @xmath286.\ ] ] now the only elements in the @xmath287^nd^ row of @xmath288 which are in @xmath242 are @xmath289=[r-3 , r-3]$ ] and @xmath290=[r-3 , r-5]$ ] .
thus by expanding along the @xmath287^nd^ row we obtain @xmath291q_1+[a_{r-2 } , b_{r-6}]q_2)$ ] with @xmath292\ ] ] and @xmath293.\ ] ] expanding all the minors we obtain in this way , step by step along the @xmath294^th^ rows , with @xmath295 , and arguing in the same way , we get @xmath296p)$ ] with @xmath297 a polynomial involving only indeterminates in @xmath242 , and @xmath298
. thus @xmath299[1,b_r]p-[2,b_r][1|c]p)$ ] which is bigger than @xmath168 since @xmath300 $ ] and @xmath301 $ ] are not in @xmath242 .
now we are in position to give a description of some leading terms of @xmath50-minors of @xmath15 with respect to @xmath25 .
[ pr : initmon ] let @xmath168 be a square - free monomial of degree @xmath50 for some @xmath64 in the set @xmath109 such that for all @xmath302 and @xmath303 with @xmath304 one has @xmath305 and @xmath306 . then @xmath168 is the leading term of an @xmath50-minor of @xmath15 with respect to @xmath25 .
let @xmath307 be a squarefree monomial with @xmath308 for every @xmath309 and @xmath310 .
we prove by induction on @xmath243 that if @xmath311 and @xmath312 for every @xmath313 , @xmath314 , then @xmath168 is the leading term of an @xmath50-minor @xmath73 of the matrix @xmath15 with respect to @xmath25 .
the induction base @xmath315 is a consequence of [ le : initial1 ] and we may now assume @xmath316 .
we set @xmath317 with @xmath318 . by the [ le :
initial1 ] there exists @xmath319 such that for @xmath320 $ ] we have @xmath321 and @xmath322 . to conclude the proof
we show that @xmath323 , with @xmath324.\ ] ] note that by assumption @xmath325 and @xmath326 for every @xmath313 , @xmath314 , thus all the row indices ( resp .
the column indices ) in @xmath73 are distinct and @xmath327 . expanding @xmath73 along its first row one has @xmath328[i_2,\ldots , i_j , a_1,\ldots , a_p\,|\,i_2,\ldots , i_j , b_1,\ldots , b_p]+ \sum_{k=2}^{j } \pm [ i_1 \,|\,i_k]m_k+\sum_{h=1}^{p } \pm [ i_1 \,|\,b_h]n_h.\ ] ] by induction @xmath329)=x_{i_2\ , i_2}\cdots x_{i_j\ , i_j}\cdot m'$ ] , thus to conclude we have to prove that in the two sums in the expansion can not appear any term bigger than @xmath168 .
first consider the terms in @xmath330m_k.$ ] let @xmath331 if @xmath332 and @xmath333 , then @xmath334 for every @xmath335 , thus no term bigger than @xmath168 appears in the sum . if @xmath336 , then @xmath337 or @xmath338 and @xmath339
. suppose @xmath337 ( in the other case one concludes similarly ) .
the only possible terms bigger than @xmath168 come from @xmath340m_2=[i_1 \,|\,i_1 + 2][i_2,\ldots , i_j , a_1,\ldots , a_p\,|\ , i_1,i_3\ldots , i_j , b_1,\ldots , b_p]$ ] .
if @xmath341 we conclude by expanding @xmath342 along its first row . otherwise we can repeat the reasoning until we find @xmath343 such that @xmath344 and we conclude .
it remains to consider the cases @xmath345 and @xmath346 , that can be treated similarly , by expanding along the first row and remembering that @xmath347 and @xmath348 are the smallest indeterminates in @xmath242 .
consider now the terms in @xmath349n_h.$ ] the only terms to be considered are the ones with @xmath350 , that is @xmath351 let start with @xmath352 , that is @xmath353 and @xmath354 .
in particular @xmath355 , then @xmath356 for every @xmath335 , thus @xmath357 and @xmath358.$ ] by developing @xmath73 along its first row one has : @xmath359[a_1,\ldots , a_p\,|\ , b_1,\ldots , b_{p-1},n ] \pm [ n-1\,|\,n][a_1,\ldots , a_p\,|\,n-1,b_1,\ldots , b_{p-1}]+ \textrm { other terms } \ ] ] all containing a variables not in @xmath242 ; thus the conclusion follows by induction and by the fact that @xmath348 is the smallest variables in @xmath242 .
similarly one concludes in the case @xmath360 .
consider now @xmath361 , that is @xmath362 .
in particular @xmath363 and @xmath364 .
note that @xmath365 does not appear in the first row of @xmath73 on the right of @xmath347 , otherwise it would be @xmath366 for some @xmath14 and we would have @xmath367 , that contradicts the hypothesis . now if @xmath368 , the conclusion follows by developing @xmath73 along the first row and noting that @xmath347 is the smallest indeterminate in @xmath369 . if @xmath370 , developing @xmath73 along the first row we have to consider the term @xmath371 $ ] . if @xmath372 the leading term of @xmath373 is divided by @xmath374 , thus it is smaller than the term we want to show to be the leading one ; so we are done .
if @xmath375 we go on expanding until we find @xmath376 and we conclude in the same way .
suppose @xmath377 , for some @xmath335 , that is , @xmath378 ; @xmath379 , otherwise it would be @xmath380 , that contradicts the hypothesis .
thus we are in one of the cases 2,3,4 of [ le : initial1 ] .
there are then only two possibilities : @xmath381 is equal to @xmath382 or to @xmath383 , thus @xmath384 which contradicts the hypothesis . or @xmath381 is equal to @xmath385 or to @xmath386 which leads to a similar contradiction .
thus this can not occur .
analogously , one proves that it can not be @xmath387 .
if @xmath388 for some @xmath389 , that is @xmath390 , then @xmath391 in fact @xmath392 would imply @xmath393 , that contradicts the hypothesis .
one concludes also in this case arguing as in the case @xmath361 , and this concludes the proof .
we introduce some notions that will be used to describe the simplicial complex associated to @xmath394 .
denote by @xmath395 the boundary complex of the @xmath396-dimensional cyclic polytope with @xmath168 vertices .
recall that , by gale s evenness condition , the facets of @xmath395 are the subsets @xmath397 $ ] of size @xmath396 such that for any two @xmath398 \setminus m$ ] with @xmath399 the number of elements @xmath400 for which @xmath401 is even ( see ( * ? ? ? * theorem 0.7 ) ) .
moreover let @xmath402,e)$ ] be the @xmath168-cycle graph on vertex set @xmath403 = \{1,\ldots , m\}$ ] and edge set @xmath404 .
for some @xmath405 consider the simplicial complex @xmath406 consisting of all subsets @xmath73 of @xmath403
$ ] such that the vertices in @xmath73 are covered by a ( partial ) matching of @xmath407 of size @xmath181 .
first we note that by definition @xmath406 is a pure simplicial complex and that the boundary complex of @xmath408 is pure as a boundary complex of a simplicial polytope .
we use gale s evenness condition . by definition
the facets of @xmath406 are given by the vertex set of a matching of size @xmath181 in @xmath407 .
if @xmath73 is such a set then for any two @xmath398 \setminus m$ ] such that @xmath399 the number @xmath401 of elements @xmath400 that lie between them must be even as they are exactly the elements covered by a set of disjoint edges .
thus by gale s evenness condition it follows that @xmath73 lies in @xmath408 .
conversely , if @xmath73 is a facet of the boundary complex of @xmath408 then between any two @xmath411 \setminus m$ ] where @xmath399 the number of @xmath401 is even .
thus by choosing @xmath412 and @xmath243 with the property that @xmath413 one sees that by a partial matching of @xmath407 one can cover all elements @xmath134 of @xmath73 for which there are @xmath398 \setminus m$ ] such that @xmath401 . since in @xmath73
we have @xmath414 vertices of which an even number is covered , an even number is left .
those remaining vertices are an initial and final segment of @xmath403 $ ] and therefore can be covered by another few edges of @xmath407 that form a partial matching . thus @xmath415 .
the set @xmath416 is a face of @xmath406 if and only if we can find @xmath181 edges that cover @xmath416 .
but @xmath416 is of size @xmath417 and does not contain any edge . hence @xmath416 can not be covered by @xmath181 edges and @xmath416 is a nonface .
now let @xmath419 be some vertex .
then @xmath420 contains @xmath181 elements .
no two elements of @xmath421 lie in an edge .
starting from any @xmath422 we go around @xmath423 in a fixed order .
we pair each element of @xmath420 with its neighbor in this order .
since no neighbor is in @xmath416 this will give @xmath181 edges covering @xmath421 .
hence @xmath420 is a face .
in particular @xmath416 is a minimal nonface .
now it remains to be shown that any minimal nonface @xmath416 of @xmath406 satisfies ( i ) and ( ii ) .
let @xmath416 be any minimal nonface of @xmath406 . by @xmath405 the full ground set
@xmath403 $ ] is not a minimal nonface .
hence , there is a vertex @xmath424 that is not contained in @xmath416 .
starting from @xmath424 we go in a fixed direction around @xmath425 .
we mark a vertex red if it is in @xmath416 and the preceding vertex is not yet marked red .
we mark a vertex blue if it is in @xmath416 and the preceding vertex is marked red .
we mark a vertex green if it is not in @xmath416 but the preceding vertex is marked red .
it follows that @xmath416 consists of all red and blue vertices .
now remove from @xmath416 all blue vertices .
then the resulting set @xmath426 does not contain any edge .
thus if @xmath426 has @xmath417 or more elements then it contains a subset satisfying ( i ) and ( ii ) .
since we know that all subsets satisfying ( i ) and ( ii ) are minimal nonfaces , it follows that @xmath416 itself must satisfy ( i ) and ( ii ) .
hence we are left with the situation when @xmath426 contains strictly less than @xmath417 vertices .
but by construction the vertex following a red vertex is either blue or green .
hence the set of red , blue and green vertices is a set containing @xmath416 and being contained in a matching of size @xmath181 .
thus @xmath416 can not be a nonface . for exhibiting the connection of the previous lemmas with @xmath394
, we consider a graph on vertex set @xmath242 , with @xmath427 the edges are formed by the two elements subsets that contain one element @xmath428 in @xmath429 and one element in @xmath87 that lies either in the same row or column as @xmath428 .
one easily sees that this graph is a @xmath38-cycle whose vertices alternate between elements in @xmath429 and elements in @xmath87 ( see figure 1 ) .
the preceding lemmas for @xmath430 and @xmath431 imply the following proposition . for the proof of the proposition we need a simple lemma that can for example
be found in ( * ? ? ?
* lemma 5.1 ) .
a version of this for arbitrary ( not necessarily monomial ) ideals is stated in ( * ? ? ?
* lemma 4.2 ) .
let us identify the variables in @xmath109 with the elements of @xmath441 $ ] as indicated in [ fig : figure1 ] .
then [ pr : initmon ] implies that the monomials whose support sets are satisfy the conditions of [ le : nonface ] lie in @xmath394 .
thus [ le : nonface ] implies that the stanley - reisner ideal of @xmath432 is a subset of @xmath394 .
the dimensions of the respective quotient rings and their multiplicities coincide by [ le : cyclic ] and [ co : n-2 ] . hence using [ monomialinclusion ]
it follows that @xmath394 equals the stanley - reisner ideal of @xmath442 .
the following is an immediate consequence of the well known fact that the stanley - reisner ideal of the boundary complex of a simplicial polytope is gorenstein ( see for example ( * ? ? ?
* corollary 5.5.6 ) ) . | we show that the ideal generated by the @xmath0 minors of a general symmetric @xmath1 by @xmath1 matrix has an initial ideal that is the stanley - reisner ideal of the boundary complex of a simplicial polytope and has the same betti numbers . |
the fundamental measure density functional theory for hard spheres is generalized to binary mixtures of arbitrary positive and moderate negative non - additivity between unlike components . in bulk
the theory predicts fluid - fluid phase separation into phases with different chemical compositions .
the location of the accompanying critical point agrees well with previous results from simulations over a broad range of non - additivities and both for symmetric and highly asymmetric size ratios .
results for partial pair correlation functions show good agreement with simulation data . density - functional theory ( dft )
is a powerful approach to study equilibrium properties of inhomogeneous systems , including dense liquids and solids of single- and multi - component substances @xcite .
its practical applicability depends on the quality of the approximation to the central object , the ( helmholtz ) excess free energy functional arising from the interparticle interactions .
the specific model of additive hard sphere mixtures constitutes the reference system par excellence to describe mixtures governed by steric repulsion , and rosenfeld s fundamental - measure theory ( fmt ) @xcite is arguably the best available approximation to tackle inhomogeneous situations .
a rapidly increasing number of applications to interesting physical problems can be witnessed @xcite .
the more general non - additive hard sphere mixture is defined through pair potentials between particles of species @xmath0 and @xmath1 , given as @xmath2 for @xmath3 and 0 otherwise , where @xmath4 is the center - center distance between the two particles , and @xmath5 is the distance of minimal approach between particles of species @xmath0 and @xmath1 . in a binary mixture non - additivity is measured conventionally through the parameter @xmath6 . the physics of non - additive hard sphere mixtures is considerably richer than that of the additive case .
in particular the case of @xmath7 is striking , as small values of @xmath8 are known to be already sufficient to induce stable fluid - fluid demixing into phases with different chemical compositions ( for recent studies see e.g. references @xcite ) .
the treatment of general non - additivity is elusive in the fmt framework .
the author is aware of successful studies only in four special cases : first , for the asakura - oosawa - vrij ( aov ) model @xcite , where species 1 represents colloidal hard spheres and species 2 ( with @xmath9 ) represents non - interacting polymer coils with radius of gyration equal to @xmath10 , an excess free energy functional was given @xcite .
second , a free energy functional for the widom - rowlinson ( wr ) model , where @xmath11 , was obtained @xcite .
third , the depletion potential between two big spheres immersed in a sea of smaller spheres was obtained through `` roth s trick '' of working on the level of the one - body direct correlation functional @xcite . in this case
the functional for the additive case is sufficient to obtain results , but the approach is limited to small concentration of big spheres .
fourth , in lafuente and cuesta s fmt for lattice hard core models , due the an odd - even effect of the particle sizes ( measured in units of lattice constants ) , non - additivity of the size of one lattice spacing arises @xcite .
this effect , however , is specific to lattice models and vanishes in the continuum limit .
the aim of the present letter is to generalize fmt for hard spheres to the case of general positive and moderate negative non - additivity and arbitrary size asymmetry .
the proposed extended framework accommodates , in the respective limits , the rosenfeld functional for additive hard sphere mixtures @xcite , the dft for the extreme non - additive aov case @xcite , and the exact virial expansion up to second order in densities .
the structure of the theory , however , goes qualitatively beyond that of either limit . between pairs of particles of species
@xmath12 and 22 are related to radii through @xmath13 , @xmath14 and @xmath15 , respectively .
the spheres of radii @xmath16 and @xmath17 represent the weight functions @xmath18 and @xmath19 , respectively , and can be viewed as `` true '' particle shapes .
the sphere of radius @xmath20 represents the kernel @xmath21 being a mere construct to generate the correct hard core distance @xmath22 between species 1 and 2 . ]
the excess ( over ideal ) helmholtz free energy functional is expressed as @xmath23 = k_bt\int d{{\bf x}}d{{\bf x } } ' \sum_{\alpha,\beta=0}^3 k^{(12)}_{\alpha\beta}(|{{\bf x}}-{{\bf x}}'| ) \phi_{\alpha\beta}\left(\{n_\nu^{(1)}({{\bf x}})\},\{n_\tau^{(2)}({{\bf x}}')\}\right ) , \label{eqfexc}\ ] ] where @xmath24 is the one - body density distributions of species @xmath25 dependent on position @xmath26 , @xmath27 is the thermal energy , @xmath28 for @xmath29 is the free energy density depending on the sets of weighted densities @xmath30 for @xmath25 , and the kernels @xmath31 are a means to control the range of non - locality between unlike components and depend solely on distance @xmath4 .
the weighted densities are built in the usual way @xcite through convolution of the respective bare density profile with appropriate weight functions : @xmath32 where @xmath33 labels the type of weight function , and @xmath34 is the particle radius of species @xmath25 .
the ( fully scalar ) kierlik - rosinberg form @xcite of the @xmath35 is used in the following , as this renders the determination of the @xmath31 more straightforward .
the @xmath35 are @xmath36 where @xmath37 , the prime denotes the derivative w.r.t . the argument , @xmath38 is the dirac distribution , and @xmath39 is the step function . alternatively , in fourier space
the weight functions are @xmath40 and given as @xmath41 with the abbreviations @xmath42 and @xmath43 .
the kernels @xmath31 in ( [ eqfexc ] ) can be viewed as @xmath44-components of a second - rank tensor @xmath45 where indexing is such that the top row contains @xmath46 , etc , and @xmath47 distinguishes different elements .
all @xmath31 possess a range of @xmath48 , i.e. vanish for values of @xmath4 beyond that distance ( see figure [ figillustration ] for an illustration of the length scales ) .
the dimension of @xmath21 is @xmath49 , and hence the dimension of @xmath50 is @xmath51 .
the elements of @xmath52 are defined , with @xmath53 , through ( [ eqwfirst ] ) , and furthermore @xmath54 with the derivatives @xmath55 for @xmath56 .
again we also give the fourier space representation [ being together with ( [ eqweightsinfourierspace ] ) also valid for @xmath57 , which reads @xmath58 in order to express the dependence of the free energy density , @xmath28 in equation ( [ eqfexc ] ) , on the weighted densities ( [ eqweighteddensities ] ) we introduce ansatz functions @xmath59 for species @xmath25 that possess the dimension of @xmath60 and the order @xmath61 in density ( i.e. contain @xmath61 factors @xmath62 ) .
explicit expressions for the non - vanishing terms are @xmath63 the excess free energy density is then constructed as @xmath64 where @xmath65 is the @xmath61th derivative of the zero - dimensional excess free energy as a function of the average occupation number @xmath66 @xcite , @xmath67 , and @xmath68 for @xmath69 .
the specific form ( [ eqphiofweighteddensities ] ) ensures both that the terms in the sum in ( [ eqfexc ] ) possesses the correct dimension of @xmath70 and that the prefactor of @xmath71 in ( [ eqphiofweighteddensities ] ) is of the total order @xmath61 in densities , as is common in fmt .
this completes the prescription for the functional ; a full account of all details , also for multi - component mixtures and for lower spatial dimensionality , will be given elsewhere .
here we discuss some of the properties of the theory . for small densities
it is straightforward to show that the correct virial expansion up to second order in densities is obtained , @xmath72 , where the mayer functions , @xmath73 , are recovered through @xmath74 where @xmath75 denotes the convolution , @xmath76 . in the limit of an additive mixture , @xmath77 and hence @xmath78 , one finds that @xmath79 if @xmath80 and @xmath81 otherwise .
this leads to a cancellation of one spatial integration in ( [ eqfexc ] ) and yields the rosenfeld functional for a binary additive hard sphere mixture @xcite with radii @xmath16 and @xmath17 . in the aov limit , @xmath82
, one finds that @xmath83 if @xmath84 , and @xmath85 otherwise .
the integration over @xmath86 in ( [ eqfexc ] ) together with the kernel @xmath87 and the fact that the density @xmath88 appears linearly in @xmath28 , see @xmath89 in ( [ eqcoefficienta01 ] ) , plays the same role that building weighted densities for the polymer species in the aov case does .
the resulting functional is equal to that for the aov model @xcite .
however , in the wr limit , in contrast to @xcite , terms higher than on the second virial level vanish . for @xmath90
the two species decouple , and @xmath91={\cal f}_{\rm exc}[\rho_1]+ { \cal f}_{\rm exc}[\rho_2]$ ] which is _ not _ obeyed by the present approximation , limiting its applicability to small values of @xmath92 .
we next turn to an investigation of bulk properties of the theory .
to assess structure , direct correlation functions can be obtained via @xmath93 which can be shown to feature percus - yevick ( py ) like behavior : @xmath94 . inverting the ornstein - zernike ( oz ) relations
permits to calculate partial structure factors , @xmath95 , and partial pair correlation functions , @xmath96 .
we have carried out monte carlo ( mc ) computer simulations in the canonical ensemble with 1024 particles and @xmath97 mc moves per particle ; histograms of all distances between particles yield benchmark results for @xmath96 .
we have chosen an intermediate size ratio of @xmath98 and have considered various values of @xmath8 from @xmath99 to 0.5 and a range of statepoints characterized by packing fractions , @xmath100 for @xmath25 . for @xmath101 ,
the current dft reproduces the solution of the py integral equation , as the functional reduces to the rosenfeld case , which is known to yield the same @xmath102 as the py approximation .
results for the representative case @xmath103 at two different statepoints are shown in figure [ figgofr ] .
the core condition , @xmath104 , is only approximately fulfilled , but the overall agreement between results from theory and simulation is quite remarkable . , between species @xmath12 and 22 ( as indicated ) , as a function of the scaled distance @xmath105 , as obtained from the present dft using the oz route ( dashed lines ) and from mc simulation ( solid lines ) .
results for @xmath106 ( @xmath107 ) are shifted upwards by one ( two ) units for clarity .
parameters are @xmath98 , @xmath108 , @xmath109 and @xmath110 ( lower ) , 0.1 ( upper ) . for comparison ,
the theoretical critical point is located at @xmath111 . ] in principle one could envisage that this approach permits to study the depletion potential , @xmath112 , between particles of species 1 being generated by the immersion into a `` sea '' of particles 2 through @xmath113 for @xmath114 , and @xmath115 .
however , for the ( relevant ) case of small size ratios ( e.g. @xmath116 , see @xcite ) already in both limits of additive hard spheres and the aov model the results are only of rather moderate accuracy , underestimating the strength of the depletion attraction @xcite , similar to results from the py approximation
. however , results from the present theory obtained through the oz route ( not shown ) cross over smoothly between the additive hard sphere case and the aov case , similar to the correct behavior @xcite .
hence one can conclude that the pair structure predicted by the current dft is similar to that of the py approximation .
this is a remarkable property , and one can anticipate test - particle calculations to yield superior results . evaluating ( [ eqfexc ] ) at constant density fields yields an analytic expression for the bulk excess free energy for fluid states , @xmath117 $ ] .
the total helmholtz free energy is then @xmath118 $ ] , where @xmath119 is the ( irrelevant ) de broglie wavelength of species @xmath0 , and @xmath120 is the system volume . via taylor expanding @xmath121 in both densities one can show that it features the exact second virial coefficients ( consistent with the correct incorporation of @xmath122 on the second virial level ) and also the exact third virial coefficients ( see e.g. @xcite ) provided @xmath123 .
the fluid - fluid demixing spinodal can be obtained from ( numerical ) solution of @xmath124 , and the location of the critical point can be determined from minimizing one of the chemical potentials , @xmath125 or @xmath126 , along the spinodal .
such results are compared in figure [ figcrit ] to those from simulations for @xmath127 , performed in the semi - grand ensemble by jagannathan and yethiraj @xcite and by gd @xcite , the latter study including a finite size analysis , for a variety of non - additivities ranging from @xmath128 . for the highly asymmetric case of @xmath129 results from gibbs ensemble simulations
were obtained by dijkstra @xcite . for
both size ratios the strong decrease of the total critical packing fraction with increasing values of @xmath8 , as well as the overall functional dependence are very well described by the theory .
however , the precise value at given @xmath8 is underestimated .
this behavior is not uncommon for mean - field like theories and is also present in the aov case . , where @xmath130 , for a non - additive binary hard sphere mixture as a function of the non - additivity parameter @xmath8 .
shown are results from the present dft ( lines ) and from simulations ( symbols ) for the symmetric case , @xmath131 , by gd @xcite ( filled squares ) and by jagannathan and yethiraj @xcite ( open circles ) , as well as for the highly asymmetric case of @xmath132 by dijkstra @xcite ( crosses ) . ]
a benefit of working on the level of the density functional is that the structure is consistent with the free energy . in figure [ figsij ]
partial structure factors are shown for a range of values of @xmath8 evaluated at the fluid - fluid critical point obtained from the free energy , and indeed @xmath133 . for @xmath134 ( as indicated ) , as a function of @xmath135 at the fluid - fluid critical point for size ratio @xmath132 and non - additivity @xmath136 .
the results for @xmath137 ( 1 ) are shifted upwards by 5 ( 10 ) units for clarity . ] in conclusion , having demonstrated the good accuracy of the predictions of the current theory for bulk fluid properties of the non - additive hard sphere mixture , we are confident that it is well suited to study interesting and relevant interfacial situations , like the structure and tension of interfaces between demixed phases , wetting at substrates @xcite and more .
note that any colloidal mixtures interacting with soft repulsive forces , as e.g.present in charge - stabilized dispersions , can be mapped ( e.g. by the barker - henderson procedure ) onto an effective non - additive hard sphere system .
hence one can anticipate experimental consequences of the structure and phase separation predicted by the present theory .
the treatment of freezing @xcite requires additional contributions to the free energy functional @xcite .
+ h. lwen , r. evans , r. blaak and k. jagannathan are thanked for useful comments .
support by the sfb tr6 of the dfg is acknowledged .
this work is part of the research program of fom , that is financially supported by the nwo . |
|
terahertz spectroscopy requires a reliable determination of the amplitude and phase of the terahertz radiation .
one established method for the generation and coherent detection of continuous - wave terahertz radiation is photomixing @xcite .
typically , terahertz radiation is generated by illuminating a biased photomixer ( the transmitter , tx ) with the optical beat of two near - infrared lasers , and coherent detection is achieved by measuring the photocurrent @xmath0 in a second photomixer ( the receiver , rx ) which is illuminated by both the optical beat and the terahertz radiation .
the accuracy of the quantity of interest , e.g. the transmission @xmath1 of a given sample , depends on the stability of the measured terahertz signal @xmath0 .
this signal in turn depends on the stabilities of the responsivities @xmath2 and @xmath3 of the two photomixers and on the stabilities of the optical powers of the two lasers at the two photomixers .
all six of these quantities are sensitive to temperature drifts . here
, we report on an efficient normalization of the terahertz signal by employing _ both _ photomixers as powermeters .
our aim is to minimize the uncertainty of @xmath0 . in order to motivate an expression describing @xmath0 , we start from the generation of the terahertz wave at the transmitter . for a constant bias voltage of @xmath4 ,
the transmitter on the one hand behaves like a photoconductive resistor , i.e. the dc photocurrent at the transmitter , @xmath5 , is given by the bias - dependent responsivity @xmath6 and the total incident optical power @xmath7 , @xmath8 where @xmath9 and @xmath10 denote the optical powers of the two lasers at the transmitter .
on the other hand , the interference between the two laser beams gives rise to a beat which yields a current oscillating at the difference frequency .
thus the transmitter emits a terahertz wave with power @xmath11 @xcite . in transmission geometry ,
the power at the receiver is given by @xmath12 .
the photocurrent in the receiver , @xmath13 , depends on the phase difference @xmath14 between the terahertz wave and the optical beat signal at the receiver .
the information on amplitude @xmath15 and phase difference @xmath14 can be separated by phase modulation with a mechanical delay stage or , in our case , a fiber stretcher @xcite .
we focus on the amplitude @xcite @xmath16 where @xmath17 denote the optical powers of the two lasers at the receiver , and @xmath3 and @xmath18 are the responsivity of the receiver and its voltage derivative , respectively . here , we have used the linear photocurrent - voltage characteristic of a photomixer ( see fig . 1 ) , thus @xmath19 with the electric field amplitude @xmath20 .
equation ( [ eqn : ithzvsr ] ) assumes perfect spatial overlap of the two lasers which is exactly the case in our setup where the superimposed light is guided in single - mode fibers .
however , if the photomixers are illuminated by a free beam , spatial overlap might have to be considered .
altogether this yields an expression for the amplitude of the terahertz signal , @xmath21 apparently , @xmath15 depends on all of the four optical powers @xmath22 , which in principle may drift independently . a drift of e.g. the optical output power of laser 1 affects @xmath9 and @xmath23 but not @xmath10 or @xmath24 , whereas an attenuation or mechanical displacement of the beam in only the transmitter branch will change @xmath9 and @xmath10 .
polarization effects can influence all components differently .
the polarization plays a role here because of the finger structure of our photomixers which more strongly reflects the component of the @xmath25 field parallel to the fingers than the perpendicular one ( see e.g. @xcite ) . in order to fully compensate the above mentioned drifts , one would have to measure all four optical power components as well as the responsivities ( or @xmath26 ) , which is not easily practicable . as a sophisticated alternative
, we propose to determine dc photocurrents in _ both _ photomixers , treating both of them like photodiodes .
we will show that this enables us to compensate for drifts of the responsivities and of the total optical powers illuminating the two photomixers . as a result
, the normalized signal is only sensitive to a drift of the relative power of the two lasers . to the best of our knowledge ,
biasing a photomixer at the receiver side has not been reported @xcite .
the transmitter and the unbiased receiver have been discussed above , see eqs .
( [ eqn : itx ] ) and ( [ eqn : ithzvsr ] ) . here
, we additionally bias the receiver photomixer with a small dc voltage @xmath27 of typically @xmath2830mv .
such a small bias does not change the linear behavior of the receiver photomixer ( see fig .
[ fig : ivcharacteristic ] ) , thus the slightly biased photomixer can still be used as a detector . for small values of @xmath27
we may write the responsivity as @xmath29 .
thus the dc photocurrent @xmath30 is given by @xmath31 typically , @xmath0 ( and @xmath5 ) is not a true dc current because the bias voltage at the transmitter is modulated e.g. in the khz range in order to apply lock - in detection of @xmath0 .
therefore , the dc photocurrent @xmath30 can be separated from @xmath0 .
even an unbiased photomixer shows a small residual photocurrent due to a photovoltage which arises from inhomogeneous illumination @xcite .
experimentally , we estimate this offset voltage to be of the order of a few millivolt .
the offset voltage itself is sensitive to drifts of the optical power and thus unsuitable for using the photomixer as a reliable powermeter .
this task requires a stable voltage , independent of the optical power .
hence , the reliability of the normalization is improved if one selects specimen with low offset voltages , employs low optical power , and applies a dc bias voltage which is significantly larger than the self - induced offset voltage of the photomixer . in short , we determine three different currents .
the photocurrent @xmath32 at the transmitter and the terahertz photocurrent @xmath33 at the receiver both oscillate at the frequency @xmath34 of the bias modulation of the transmitter , whereas @xmath30 is a true dc current due to the dc bias voltage at the receiver .
then , @xmath5 and @xmath30 are used to normalize @xmath0 , @xmath35 where @xmath36 and @xmath37 are constants , e.g. the long - term average values of @xmath5 and @xmath30 , respectively .
note that the normalized photocurrent does not depend on the responsivities @xmath2 and @xmath3 anymore .
this is valid under the assumption that @xmath38 is equal for a constant @xmath27 and for applying a terahertz field @xmath39 , which is supported by our results described below .
let us now examine how this normalization affects the sensitivity to a drift of the optical power .
the partial derivatives of @xmath0 and @xmath40 with respect to e.g. @xmath9 and @xmath10 are given by @xmath41 @xmath42 compared to @xmath0 , we find that @xmath40 is less sensitive to a variation of @xmath9 by a factor of @xmath43 in the limit of @xmath44 , a small drift of @xmath9 is fully compensated .
moreover , @xmath40 also stays constant for a common attenuation of both lasers , i.e. for power changes @xmath45
. however , full compensation is not achieved if @xmath9 and @xmath10 are significantly different and , at the same time , change differently . therefore , equal laser powers are preferable for the normalization .
an analogous calculation yields the same results for the receiver .
it is instructive to rewrite @xmath40 in terms of the splitting ratios of the two laser powers at the transmitter @xmath46 and at the receiver @xmath47 , @xmath48^{1/2 } \ , [ r_\mathrm{rx}(1-r_\mathrm{rx})]^{1/2 } \ , .
\label{eqn : ithznormsplittingratio}\ ] ] the normalized terahertz photocurrent is no longer a function of the total power or of the responsivities of the photomixers , but it only depends on the splitting ratios in the two branches . in the simple case of equal splitting ratios for transmitter and receiver , @xmath49 , eq .
( [ eqn : ithznormsplittingratio ] ) reduces to @xmath50
a sketch of our experimental setup is given in fig .
[ fig : setup ] , for details we refer to refs . @xcite . for both seed lasers , the optical power at a given frequency
is actively stabilized to about @xmath511% , but the tapered amplifier and the photomixers are sensitive to temperature drifts .
the photocurrent at the receiver is preamplified and then digitized to determine the terahertz photocurrent @xmath0 and the dc receiver photocurrent @xmath30 .
the transmitter photocurrent @xmath5 is measured with the help of a 1k@xmath52 resistor in series with the photomixer .
the signals of both optical frequencies are amplified simultaneously in the tapered amplifier , and the resulting beam with a particular value of @xmath53 is sent to both photomixers . thus the simple case of equal splitting ratios ( see eq .
( [ eqn : r ] ) ) is applicable to our setup .
however , small changes of @xmath54 and @xmath55 in the path from the amplifier to the two photomixers can not be excluded .
the gain of the tapered amplifier is wavelength dependent due to reflections at the surfaces of the amplifier chip .
a temperature - induced change of the chip length therefore leads to a drift of the ratio of the laser powers which can typically range from 40:60 to 50:50 , equivalent to @xmath56 to 0.5 .
we thus expect that fluctuations of the normalized photocurrent @xmath40 are suppressed to below 5% .
as a first example we study the effect of an artificially introduced drift @xmath57 of the power @xmath58 of one laser ( @xmath59 and @xmath60 ) . for this measurement , the tapered amplifier was removed from the setup .
note that @xmath58 was not measured directly at the photomixers but with a separate photodiode .
experimentally , we started with equal laser powers , @xmath61 , and then reduced @xmath58 , keeping the power of the second laser fixed ( see fig .
[ fig : normpc ] ) . according to eqs .
( [ eqn : ithzvsefield ] ) and ( [ eqn : r ] ) , we expect that the terahertz photocurrent @xmath0 shows the same relative drift as @xmath58 , i.e. @xmath62 , whereas the relative drift of the normalized current @xmath40 is expected to be much smaller ( solid lines in fig .
[ fig : normpc ] ) .
this is corroborated by the experimental data .
however , the measured drift @xmath63 is slightly larger than expected .
this is mainly due to the fact that @xmath64 shows additional fluctuations of the order of roughly 3% .
if we subtract these deviations between @xmath64 and the straight line from @xmath65 , then the result is very close to the predicted curve ( see crosses in fig . [
fig : normpc ] ) .
the remaining difference can be attributed to , e.g. , the small power - dependent offset voltage at the receiver caused by inhomogeneous illumination .
although the measured data do not fully reach the ideal case of the theoretical prediction , the stability of the signal is significantly enhanced by the normalization .
for instance a reduction of @xmath58 of 20% gives rise to only 3% change of the measured value of @xmath40 . in the second example , see fig .
[ fig : splittingratio ] , we changed the splitting ratio @xmath53 while keeping the total laser power @xmath66 fixed . in this case , @xmath40 follows the predicted behavior @xmath67 excellently . as a third example
, we deployed the complete spectrometer from fig .
[ fig : setup ] and monitored the terahertz amplitude at 600ghz over 2 hours while periodically changing the laboratory temperature between 22@xmath68c and 24@xmath68c ( see fig .
[ fig : data ] ) .
the measured amplitude varies by approximately 15% peak - to - peak , whereas the stability of the normalized amplitude is improved to about 4% peak - to - peak ( bottom panel ) . as discussed in section [ sec : setup ] , in our setup the splitting ratio @xmath53 drifts between 0.4 and 0.5 as a function of temperature . according to eq .
( [ eqn : r ] ) , this corresponds to a drift of @xmath40 of 4% , in excellent agreement with our data .
the dc photocurrents in the transmitter and receiver are correlated to some extent ( see fig . [
fig : data ] , second and third panel ) , which here is mainly due to drifts of the tapered amplifier which is common to both optical paths .
but there are also significant differences in the fluctuations of @xmath5 and @xmath30 , substantiating the necessity to measure both currents separately .
the normalized amplitude is less stable if only @xmath5 or only @xmath30 is used for the normalization .
this is evident from the fourth panel of fig .
[ fig : data ] , which shows @xmath69 and @xmath70 . in the latter , the factor
@xmath71 has been replaced by @xmath72 in eq .
( [ eqn : ithznorm ] ) . in fig .
[ fig:100percentline ] we show the 100% line of the spectrometer in the frequency range from 50ghz to 800ghz .
the measurements were performed directly after switching on the setup . for a step size of 1ghz ,
a single run took about 10 minutes , and there was a delay of about 10 minutes between the two runs .
the initial change of the temperature gives rise to a drift of the 100% line , which is significantly reduced by the normalization .
as discussed above for the data of fig .
[ fig : data ] , the normalization is very well suited for temperature - induced drifts .
however , it does not have any effect on the noise - like features present in the 100% line .
these are caused by e.g. the uncertainty in the determination of the amplitude @xmath15 from the raw data measured as a function of @xmath14 ( see eq .
( [ eqn : ithzvsefield ] ) ) and by small fluctuations of the frequency .
finally , we varied the receiver bias and measured the noise photocurrent , i.e. the standard deviation of the terahertz photocurrent with blocked terahertz beam .
we found that the noise photocurrent depends on the details of the photomixer device .
for some devices , the noise is nearly independent of the bias , while other devices show a significant increase of the noise photocurrent when a bias of a few 10mv is applied .
presumably , this difference originates in differences in the photomixer resistivity . in order to obtain stable and reliable data ,
it is of course desirable to reduce the fluctuations of the laboratory temperature in the first place .
however , stabilization to much better than @xmath73 is not a trivial task .
moreover , the stabilization achieved via photocurrent normalization may be a significant improvement for terahertz applications requiring measurements outside a regular laboratory .
as an alternative to the method described here , one may consider to monitor the optical power of the two lasers with , e.g. , a photodiode .
however , the discussed normalization via the photocurrents has two main advantages .
it does not require an extra sensor , and the power is measured directly within the photomixers .
therefore , the normalization also compensates drifts of the responsivity of the photomixers , e.g. caused by mechanical drifts within the photomixer device .
we have described and demonstrated a normalization scheme for the terahertz photocurrent in a continuous - wave photomixing system .
this method is based on measuring the dc photocurrents in both the transmitter and receiver photomixers .
consequently , no extra sensor is needed and the method can easily be implemented .
any change of the laser power illuminating the photomixers can be described as a change of the total laser power in combination with a changing splitting ratio @xmath53 of the two laser powers .
the normalization fully compensates drifts of the total laser power as well as drifts of the responsivities , thus the stability of the normalized terahertz signal only depends on the splitting ratio @xmath53 .
we have provoked large changes of the terahertz signal by either reducing the power of one laser or by changing the laboratory temperature to simulate unstable ambient conditions . in all cases ,
the normalized signal is stable within a few percent .
this project is supported by the dfg via sfb 608
. 99 k.a .
mcintosh , e.r .
brown , k.b .
nichols , o.b .
mcmahon , w.f .
dinatale , and t. m. lyszczarz , `` terahertz photomixing with diode lasers in low - temperature - grown gaas , '' appl .
. lett . * 67 * , 3844 ( 1995 ) .
s. verghese , k.a .
mcintosh , s. calawa , w.f .
dinatale , e.k .
duerr , and k.a .
molvar , `` generation and detection of coherent terahertz waves using two photomixers , '' appl .
. lett . * 73 * , 3824 - 3826 ( 1998 ) .
deninger , t. gbel , d. schnherr , t. kinder , a. roggenbuck , m. kberle , f. lison , t. mller - wirts , and p. meissner , `` precisely tunable continuous - wave terahertz source with interferometric frequency control , '' rev .
instr . * 79 * , 044702 ( 2008 ) .
a. roggenbuck , h. schmitz , a. deninger , i. cmara mayorga , j. hemberger , r. gsten , and m grninger , `` coherent broadband continuous - wave terahertz spectroscopy on solid - state samples , '' new j. phys . *
* 1**2 , 043017 ( 2010 ) .
s. matsuura , h. ito , `` generation of cw terahertz radiation with photomixing , '' in _ terahertz optoelectronics , _
k. sakai , ed .
( springer - verlag berlin heidelberg , 2005 ) 157 - 202 .
d. saeedkia , s. safavi - naeini , `` terahertz photonics : optoelectronic techniques for generation and detection of terahertz waves , '' j. lightwave technol .
* 26 * 2409 - 2423 ( 2008 ) .
s. preu , g.h .
dhler , s. malzer , l.j .
wang , a.c .
gossard , `` tunable , continuous - wave terahertz photomixer sources and applications , '' j. appl . phys . * 109 * 61301 - 61356 ( 2011 ) .
bjarnason and e.r .
brown , `` sensitivity measurement and analysis of an eras : gaas coherent photomixing transceiver , '' appl .
. lett . * 87 * , 134105 ( 2005 ) .
a. roggenbuck , k. thirunavukkuarasu , h. schmitz , j. marx , a. deninger , i. cmara mayorga , r. gsten , j. hemberger , and m. grninger , `` using a fiber stretcher as a fast phase modulator in a continuous wave terahertz spectrometer , '' j. opt .
b * 29 * , 614 - 620 ( 2012 ) .
i. cmara mayorga , e.a .
michael , a. schmitz , p. van der wal , r. gsten , k. maier , and a. dewald , `` terahertz photomixing in high energy oxygen- and nitrogen - ion - implanted gaas , '' appl .
. lett . * 91 * , 031107 ( 2007 ) .
biased uni - travelling - carrier photodiodes have been employed as receivers , see t. nagatsuma , a. kaino , s. hisatake , k. ajito , h .- j .
song , a. wakatsuki , y. muramoto , n. kukutsu , and y. kado , piers online * 6 * , 390 - 394 ( 2010 ) , but there the purpose of biasing is not related to normalization .
brown , k.a .
mcintosh , f.w .
smith , k.b .
nichols , m.j .
manfra , c.l .
dennis , and j.p .
mattia , `` milliwatt output levels and superquadratic bias dependence in a low - temperature - grown gaas photomixer , '' appl . phys .
* 64 * , 3311 - 3313 ( 1994 ) .
fig . 1 .
current - voltage characteristic of an illuminated photomixer . at low voltages ,
the photocurrent is proportional to the voltage ( black points : data , red line : linear fit ) .
consequently , the photomixer converts a terahertz electric field linearly into a photocurrent ( see eq .
( [ eqn : ithzvsefield ] ) ) and can be used as a detector .
inset : photocurrent - voltage characteristic for larger voltages ( see also @xcite ) .
drift of the measured terahertz photocurrent @xmath0 ( black ) and of the normalized photocurrent @xmath40 ( red ) upon reduction of the power @xmath58 of one of the two lasers .
@xmath74 corresponds to equal laser powers , i.e. , @xmath61 or @xmath75 .
solid lines : predictions according to eqs .
( [ eqn : ithzvsefield ] ) and ( [ eqn : r ] ) .
full symbols : measured data .
crosses are obtained by subtracting the deviation between black symbols and black line from the red symbols .
fig . 4 .
normalized terahertz photocurrent @xmath40 vs. splitting ratio of the two laser powers , @xmath76 , for a constant total power @xmath77 .
the solid line depicts our expectation @xmath78 , see eq .
( [ eqn : r ] ) .
@xmath40 is scaled such that @xmath79 .
effect of a periodic variation of the laboratory temperature by @xmath80k ( top panel ) on the stability of the measured terahertz signal @xmath0 ( black line , bottom panel ) .
the drift is strongly suppressed in the normalized photocurrent @xmath40 ( red line , bottom panel ) .
second and third panel : dc photocurrents @xmath5 and @xmath30 at transmitter and receiver , respectively .
fourth panel : normalized photocurrent if only @xmath5 ( green ) or only @xmath30 ( blue ) is used for the normalization , e.g. @xmath81 .
( black ) and of the normalized photocurrent @xmath40 ( red ) upon reduction of the power @xmath58 of one of the two lasers .
@xmath74 corresponds to equal laser powers , i.e. , @xmath61 or @xmath75 .
solid lines : predictions according to eqs .
( [ eqn : ithzvsefield ] ) and ( [ eqn : r ] ) .
full symbols : measured data .
crosses are obtained by subtracting the deviation between black symbols and black line from the red symbols .
fig3normpc.eps . ] vs. splitting ratio of the two laser powers , @xmath76 , for a constant total power @xmath77 .
the solid line depicts our expectation @xmath78 , see eq .
( [ eqn : r ] ) .
@xmath40 is scaled such that @xmath79 .
fig4ithzvssplittingratio.eps . ]
k ( top panel ) on the stability of the measured terahertz signal @xmath0 ( black line , bottom panel ) .
the drift is strongly suppressed in the normalized photocurrent @xmath40 ( red line , bottom panel ) .
second and third panel : dc photocurrents @xmath5 and @xmath30 at transmitter and receiver , respectively .
fourth panel : normalized photocurrent if only @xmath5 ( green ) or only @xmath30 ( blue ) is used for the normalization , e.g. @xmath81 . fig5photocurrentcorrectionfigpaper.eps . ] | in a continuous - wave terahertz system based on photomixing , the measured amplitude of the terahertz signal shows an uncertainty due to drifts of the responsivities of the photomixers and of the optical power illuminating the photomixers .
we report on a simple method to substantially reduce this uncertainty . by normalizing the amplitude to the dc photocurrents in both the transmitter and receiver photomixers
, we achieve a significant increase of the stability .
if , e.g. , the optical power of one laser is reduced by 10% , the normalized signal is expected to change by only 0.3% , i.e. , less than the typical uncertainty due to short - term fluctuations .
this stabilization can be particularly valuable for terahertz applications in non - ideal environmental conditions outside of a temperature - stabilized laboratory . |
[ the material in appendices a and b appeared in the early arxiv versions of this paper . ]
we consider binary - state dynamics on static , undirected , connected networks in the limit of infinite network size ( i.e. , @xmath11 , where @xmath12 is the number of nodes in the network )
. for convenience , we call the two possible node states _ susceptible _ and _ infected _ , as is common in disease - spread models . however , our approach also applies to other binary - state dynamics , such as spin systems @xcite , where each node may be in the @xmath13 ( spin - up ) or the @xmath14 ( spin - down ) state .
the networks have degree distribution @xmath15 and are generated by the configuration model @xcite .
dynamics are stochastic , and are defined by infection and recovery probabilities which depend on the degree @xmath0 of a node , and on the current number @xmath1 of infected neighbors of the node . thus @xmath16 is defined as the probability that a @xmath0-degree node that is susceptible at time @xmath17 , with @xmath1 infected neighbors , changes its state to infected by time @xmath18 , where @xmath19 in an infinitesimally small time interval .
similarly , @xmath20 is the probability that a @xmath0-degree infected node with @xmath1 infected neighbor moves to the susceptible state within a time @xmath19
. these general infection and recovery probabilities can describe many dynamical processes on networks .
for example , in the susceptible - infected - susceptible ( sis ) model of disease spread ( the contact process ) @xcite , each susceptible node may be infected at a rate @xmath2 by each infected neighbor , and each infected node recovers at a constant rate @xmath3 , so the rates @xmath9 and @xmath10 take the form @xmath21 zero - temperature glauber dynamics for a spin system on a network @xcite provide another example . here
, each node has a spin of @xmath13 or @xmath14 ( which we can identify with the infected or susceptible state , respectively ) . in each infinitesimal time step ,
one node is selected at random ( @xmath22 ) and its spin is set to @xmath13 if the local field ( sum of its neighbors spins ) is positive , to @xmath14 is the local field is negative , and to @xmath23 with equal probability if the local field is zero .
thus , the new spin matches the majority of its neighbors spins ( with a random choice in case of a tie ) .
the infection and recovery rates may therefore be expressed as @xmath24 the voter model @xcite has rates given by @xmath25 since infection of a @xmath0-degree node , for example , occurs by copying one of @xmath1 infected neighbors out of @xmath0 possible choices .
we now proceed to derive the master equations for dynamics of this type , closely following the approach used in @xcite for sis dynamics .
let @xmath26 ( resp .
@xmath27 ) be the set of nodes which are susceptible ( resp .
infected ) , have degree @xmath0 , and have @xmath1 infected neighbors . to quantify the size of these sets , define @xmath28 ( resp .
@xmath29 ) as the fraction of @xmath0-degree nodes that are susceptible ( resp .
infected ) at time @xmath17 , and have @xmath1 infected neighbors .
then the fraction @xmath30 of @xmath0-degree nodes that are infected at time @xmath17 is given by @xmath31 and the fraction of infected nodes in the whole network is found by summing over all @xmath0-classes : @xmath32 if a randomly - chosen fraction @xmath33 of nodes are initially infected , then the initial conditions for @xmath34 and @xmath35 are easily seen to be @xmath36 where we introduce the convenient notation @xmath37 for the binomial factor @xmath38 .
note that we can also calculate the number of edges of various types using this formalism .
for example , the number of edges in the network which join a susceptible node to an infected node ( we call these _ @xmath39-@xmath40 edges _ for short ) can be expressed in two equivalent ways : @xmath41 the first of these expressions , for example , follows from noting that in a sufficiently large network that there are @xmath42 nodes of degree @xmath0 , of which a fraction @xmath34 are susceptible and have @xmath1 infected neighbors .
each such node contributes @xmath1 edges to the total number of @xmath39-@xmath40 edges .
similar expressions may also be given for the number of @xmath39-@xmath39 and @xmath40-@xmath40 edges in the network .
we note that the equivalence of the two expressions in ( [ star ] ) is preserved by the evolution equations described below .
next , we examine how the size of the @xmath26 set changes in time .
we write the general expression @xmath43 to reflect all the transitions whose rate is linear in @xmath19 ( all other state - transitions are negligible in the @xmath44 limit ) , see fig .
[ fig_schematic ] .
here @xmath45 , for example , is the probability that a node in the @xmath26 set at time @xmath17 moves to the @xmath27 set by time @xmath18 .
it is clear from the definitions above that @xmath46 a node moves from the @xmath47 set to the @xmath26 set if it remains susceptible , while one of its susceptible neighbors becomes infected .
note this means that an @xmath39-@xmath39 edge changes to an @xmath39-@xmath40 edge .
if we suppose that @xmath39-@xmath39 edges change to @xmath39-@xmath40 edges at a ( time - dependent ) rate @xmath48 , we can write is the same for all @xmath39-@xmath39 edges in the network , regardless of their local neighborhood the same assumption is made for the other rates @xmath49 , @xmath50 , and @xmath51 .
see also the explanation in @xcite for the sis case . ]
@xmath52 since nodes in the @xmath26 set have @xmath53 susceptible neighbors , while those in the @xmath47 set have @xmath54 susceptible neighbors . to calculate @xmath48
, we count the number of @xmath39-@xmath39 edges in the network at time @xmath17 , and then count the number of edges which switch from being @xmath39-@xmath39 edges to @xmath39-@xmath40 edges in the time interval @xmath19 ; the probability @xmath55 is given by taking the ratio of the latter to the former , i.e. @xmath56 a similar approximation is used to define @xmath49 , the rate at which @xmath39-@xmath40 edges change to @xmath39-@xmath39 edges due to the recovery of an infected node : @xmath57 and we then write @xmath58 taking the limit @xmath59 of equation ( [ weqn ] ) gives the master equation for the evolution of @xmath28 ( see fig . [ fig_schematic ] ) : @xmath60 where @xmath1 is in the range @xmath61 for each @xmath0-class in the network ( and adopting the convention @xmath62 . applying identical arguments , _ mutatis mutandis _
, to the set @xmath27 , we derive the corresponding system of equations for @xmath29 : @xmath63 for @xmath64 and for each @xmath0-class in the network , with time - dependent rates @xmath50 and @xmath51 defined though @xmath34 and @xmath35 as @xmath65 the master equations ( [ seqnsapp ] ) and ( [ ieqnsapp ] ) , with the time - dependent rates @xmath48 , @xmath49 , @xmath50 and @xmath51 ( defined as nonlinear functions of @xmath34 and @xmath35 ) , form a closed system of deterministic equations which , along with initial conditions ( [ ics ] ) , can be solved numerically using standard methods .
note that the evolution equations are completely prescribed by the functions @xmath9 and @xmath10 , and so this method can be applied to any stochastic dynamical process defined by transition rates @xmath9 and @xmath10 . for the sis model , equations ( [ seqnsapp ] ) and ( [ ieqnsapp ] ) were derived in @xcite ( see also @xcite ) , with additional terms to study adaptive rewiring of the network .
the pair approximation ( [ pa ] ) derived in the main text is of the type dubbed `` heterogeneous pa '' in @xcite , because the system includes variables @xmath66 and @xmath67 for each degree class @xmath0 .
a more parsimonious set of equations may be derived under the assumptions of `` homogeneous pa '' , wherein the @xmath0-dependence of edge - based variables @xmath66 and @xmath67 is neglected . as discussed in @xcite ,
the reduction in the number of variables typically comes at the cost of reduced accuracy . for homogeneous pa , the parameter @xmath68 ( resp .
@xmath69 ) is defined as the probability that a randomly - chosen neighbor of a susceptible ( resp .
infected ) node is infected . noting that @xmath68 can be expressed in terms of @xmath34 as @xmath70 an evolution equation for @xmath71 may be derived by multiplying equation ( [ seqnsapp ] ) by @xmath72 and summing over @xmath1 and @xmath0 .
the right - hand - side of the resulting equation contains higher moments of @xmath34 , so a closure approximation is needed to proceed . if , similar to the steps yielding equations ( [ pa ] ) , we here make the ansatz that @xmath34 and @xmath35 are proportional to binomial distributions : @xmath73 then an equation for @xmath74 may be found in terms of only @xmath75 , @xmath71 , and @xmath76 .
applying the same ansatz to expression ( [ star ] ) gives the algebraic relation @xmath77 between @xmath71 and @xmath76 , where @xmath78 can be interpreted as the probability that the node at one end of a randomly - chosen edge is infected .
after some algebra , we obtain the _ homogeneous pair approximation _ , consisting of the @xmath79 differential equations : @xmath80 along with the algebraic relation ( [ algebraic ] ) , and initial conditions @xmath81 . for the sis model ,
the homogeneous pa equations ( [ hompa ] ) are identical to those recently derived by house and keeling @xcite , while for the voter model equations ( [ hompa ] ) are equivalent to those in @xcite ( in the @xmath82 limit ) .
note that on @xmath83-regular graphs , the heterogeneous and homogeneous pair approximations are identical .
99 a. barrat , m. barthlemy , a. vespignani , _ dynamical processes on complex networks _ , cambridge university press , cambridge , 2008 .
m. e. j. newman , _ networks : an introduction _
( oxford university press , oxford , 2010 ) .
r. pastor - satorras and a. vespignani , _ phys .
lett . _ * 86 * , 3200 ( 2001 ) . c. castellano , s. fortunato , and v. loreto , _ rev .
_ , * 81 * , 591 ( 2009 ) .
c. j. honey _
et al . _ ,
usa * 106 * , 2035 ( 2009 ) ; a. v. goltsev _ et al .
_ , _ phys .
e _ , * 81 * , 061921 ( 2010 ) .
a. g. haldane and r. m. may , _ nature _ , * 469 * , 351 ( 2011 ) ; r. m. may and n. arinaminpathy , _ j. r. soc . interface
, * 7 * , 823 ( 2010 ) .
s. n. dorogovtsev , a. v. goltsev , and j. f. f. mendes , _ phys .
, 66 , 016104 ( 2002 ) ; m. leone _ et al .
. j. phys .
b _ , * 28 * , 191 ( 2002 ) .
r. parshani , s. carmi , and s. havlin , _ phys . rev
_ , * 104 * , 258701 ( 2010 ) .
c. castellano and r. pastor - satorras , _ phys .
_ , * 105 * , 218701 ( 2010 ) ; r. durrett , _ proc .
usa _ , * 107 * 16413 ( 2010 ) .
s. n. dorogovtsev , a. v. goltsev , and j. f. f. mendes , _ rev .
_ , * 80 * , 1275 ( 2008 ) . k. t. d. eames and m. j. keeling , _ proc .
, * 99 * , 13330 ( 2002 ) .
e. pugliese and c. castellano , _ europhys .
_ , * 88 * , 58004 ( 2009 ) . r. m. anderson and r. m. may , _ infectious diseases of humans : dynamics and control _
( oxford university press , oxford , 1992 ) ; n. t. j. bailey , _ the mathematical theory of infectious diseases _ ( griffin , london , 1975 ) ; t. e. harris , _ ann .
_ , * 2 * , 969 ( 1974 ) .
t. m. liggett , _ interacting particle systems _ ( springer , new york , 1985 ) .
r. j. glauber , _
_ , * 4 * , 294 ( 1963 ) . m. j. de oliveira , _ j. stat . phys .
_ , * 66 * , 273 ( 1992 ) .
m. j. de oliveira , j. f. f. mendes , and m. a. santos , _ j. phys .
_ , * 26 * , 2317 ( 1993 ) .
v. marceau _ et al .
_ , _ phys .
rev . e_. * 82 * , 036116 ( 2010 ) .
s. a. levin and r. durrett , _ phil .
b _ , * 351 * , 1615 ( 1996 ) . c. castellano _ et al . _ , _ phys .
e _ , * 71 * , 066107 ( 2005 ) . c. castellano and r. pastor - satorras , _ j. stat .
_ , p05001 ( 2006 ) .
d. j. watts , _ proc .
usa _ , * 99 * , 5766 ( 2002 ) ; d. centola , v. m. eguluz , and m. w. macy , _ physica a _ , * 374 * , 449 ( 2007 ) . s. n. dorogovtsev , a. v. goltsev , and j. f. f. mendes , _ phys .
_ , * 96 * , 040601 ( 2006 ) ; a. v. goltsev , s. n. dorogovtsev , and j. f. f. mendes , _ phys
e _ , * 73 * , 056101 ( 2006 )
. g. j. baxter _ et al .
_ , _ phys .
e _ , * 82 * , 011103 ( 2010 ) .
j. p. gleeson , _ phys .
e _ , * 77 * , 046117 ( 2008 ) .
v. sood and s. redner , _ phys .
_ , * 94 * , 178701 ( 2005 ) . f. vazquez and v. m. eguluz , _ new .
_ , * 10 * , 063011 ( 2008 )
. t. house and m. j. keeling , _ j. r. soc . interface _ , * 8 * , 67 ( 2011 ) .
r. pemantle , _ ann .
_ , * 20 * , 2089 ( 1992 ) ; t. m. liggett , _ ann . probab . _ , * 24 * , 1675 ( 1996 ) . l. f. c. pereira and f. g. b. moreira , _ phys .
e _ , * 71 * , 016123 ( 2005 )
. b. a. prakash _ et al .
_ , arxiv:1004.0060 ( 2010 ) .
nol , b. davoudi , r. c. brunham , l. j. dub , and b. pourbohloul , _ phys .
e _ , * 79 * , 026101 ( 2009 ) .
f. ball and p. neal , _ math
. biosciences _ , * 212 * , 69 ( 2008 ) . v. marceau , p .- a
. nol , l. hbert - dufresne , a. allard , and l. j. dub , submitted , arxiv:1103:4059 . t. tom and r. m. ziff , _ phys .
e _ , * 82 * , 051921 ( 2010 ) .
a. v. goltsev , f. v. de abreu , s. n. dorogovtsev , and j. f. f. mendes , _ phys .
e _ , * 81 * , 061921 ( 2010 ) . s. gmez , a. arenas , j. borge - holthoefer , s. meloni , and y. moreno , _ europhys .
_ , * 89 * , 38009 ( 2010 ) . m. e. j. newman , _ phys . rev
_ , * 103 * , 058701 ( 2009 ) . j. c. miller , _ phys . rev .
e _ , * 80 * , 020901(r ) ( 2009 ) . j. p. gleeson , _ phys .
e _ , * 80 * , 036107 ( 2009 ) .
l. hbert - dufresne , p .- a .
nol , v. marceau , a. allard , and l. j. dub , _ phys .
e _ , * 82 * , 036115 ( 2010 ) .
a. hackett , s. melnik , and j. p. gleeson , _ phys .
e _ , * 83 * , 056107 ( 2011 ) .
e. a. bender and e. r. canfield , _
theory ser .
a _ , * 24 * , 298 ( 1978 ) .
b. bollobs , _ eur . j. comb .
_ , * 1 * , 311 ( 1980 ) . | binary - state dynamics ( such as the susceptible - infected - susceptible ( sis ) model of disease spread , or glauber spin dynamics ) on random networks are accurately approximated using master equations .
standard mean - field and pairwise theories are shown to result from seeking approximate solutions of the master equations .
applications to the calculation of sis epidemic thresholds and critical points of non - equilibrium spin models are also demonstrated .
.infection and recovery rates for some examples of binary - state dynamics on networks : @xmath0 is the node s degree , @xmath1 is its number of infected neighbors . parameters @xmath2 and @xmath3 are sis transmission and recovery rates ; @xmath4 and @xmath5 are the temperature and interaction strength for the ising model ; @xmath6 is the majority - vote noise parameter .
note @xmath7 glauber dynamics are identical to those of the @xmath8 majority - vote model . [ cols="^,^,^",options="header " , ] in summary , we have derived the master equations ( [ seqns])([ieqns])first introduced for sis dynamics in @xcite for general binary - state dynamics on networks , and demonstrated that their accuracy supersedes standard mf and pa methods .
mean - field and pairwise theories are derived as approximate solutions of the master equations , and equations ( [ pa ] ) explicitly give pair approximations for any dynamics defined by infection and recovery rates @xmath9 and @xmath10 .
finally , we demonstrated the application of the master equations to calculating epidemic thresholds and critical parameter values via linear stability analysis , improving significantly on existing mf and pa estimates .
we anticipate further applications of the master equation approach to the calculation of critical points in opinion models and spin systems , and expect possible extensions to include multiple - state dynamics ( such as the sir disease - spread model @xcite ) , multiple node types @xcite , discrete - time dynamics @xcite , and network models with non - zero clustering @xcite .
this work was funded by science foundation ireland awards 06/in.1/i366 and macsi 06/mi/005 .
helpful discussions with sergey melnik , rick durrett , and claudio castellano , and participants at the samsi dynamics on networks workshop are gratefully acknowledged . |
the measurement of longitudinal spin asymmetries in production of @xmath0 bosons at the relativistic heavy ion collider ( rhic ) will provide an essential probe for spin - dependent quark distributions at high momentum scales @xmath4 @xcite . at a @xmath5 center - of - mass energy @xmath6 gev ,
about @xmath7 @xmath8and @xmath9 bosons will be produced by the time the integrated luminosity reaches @xmath10 . due to the parity violation in the @xmath11 coupling , this process permits non - vanishing single - spin asymmetries @xmath12 , defined for any kinematical variable @xmath13 as@xmath14 the born - level expression for the asymmetry @xmath15 with respect to the rapidity @xmath16 of the @xmath1 boson is particularly simple if the absolute value of @xmath16 is large . in that case
, the born - level @xmath15 reduces to the ratio @xmath17 of the polarized and unpolarized parton distribution functions @xcite .
furthermore , @xmath15 tests the flavor dependence of quark polarizations .
the original method for extracting the spin - dependent quark distributions out of the rhic @xmath1 boson data is based on the direct reconstruction of the asymmetry @xmath15 @xcite .
unfortunately , such reconstruction is obstructed by specifics of the detection of @xmath0 bosons at rhic .
first , neither @xcite nor @xcite detector at rhic is hermetic .
therefore , it is not possible at those detectors to monitor the energy balance in particle reactions , even in the transverse direction with respect to the beams . due to the lack of information about the missing energy carried by the neutrino , the determination of @xmath16 is in general ambiguous and depends on the assumptions about the dynamics of the process .
second , due to the correlation between the spins of the initial - state quarks and final - state leptons , the measured value of @xmath15 is strongly sensitive to the experimental cuts imposed on the observed charged lepton .
hence , unless a theory calculation exists to reliably predict the distributions of the leptons from the @xmath1 boson decay , the spin - dependent quark distribution functions can not be determined with acceptable accuracy . in weak boson production at the fermilab tevatron @xmath18 collider ,
the mass @xmath19 of the @xmath1 boson has been measured to a great precision @xcite .
the information on @xmath19 is extracted out of the lepton - level distributions @xmath20 , @xmath21 , or @xmath22 , where @xmath23 and @xmath24 are the transverse momenta of the electron and neutrino , and @xmath25 is the transverse mass of the lepton pair . the value of @xmath19 can be determined from the analysis of the position and shape of the jacobian peaks in each of these three distributions . for most of the events contributing near the jacobian peak , the transverse momentum @xmath26 of the @xmath1 boson is very small , of order of a few gev .
therefore , at each order of the perturbative qcd calculation , there are large logarithmic contributions of the form @xmath27 , where @xmath28 and @xmath29 .
these large logarithmic contributions can be summed to all orders in @xmath30 by applying the transverse momentum resummation formalism formulated in the impact parameter space by collins , soper , and sterman ( css ) @xcite .
without such a resummation calculation performed at the decay lepton level @xcite , it would not be possible to determine @xmath19 from the experimental data .
this is because at hadron colliders the longitudinal momentum of the neutrino from the @xmath1 boson decay is not observed by the detectors ; hence , the invariant mass of the @xmath1 boson can not be measured directly . instead , the determination of the @xmath1 boson mass relies on the analysis of the shape of the jacobian peaks in the lepton - level distributions sensitive to @xmath26 .
again , without a resummation calculation at the secondary lepton level , it would not be possible to predict the shape of these distributions accurately enough to determine @xmath19 .
we refer the reader to ref .
@xcite for a detailed discussion of the phenomenology of the @xmath1 boson physics at the tevatron collider . similarly , at rhic
, a resummation calculation is needed to reliably predict the distribution of the secondary leptons from the @xmath1 boson decay .
as we already mentioned , the reason is the necessity to deduce spin - dependent parton luminosities based on the observation of the secondary charged lepton only . for rhic purposes ,
the resummation calculation at the lepton level has to be modified to include the spin dependence , which has been performed in ref .
@xcite . in this paper
, we will apply the calculation in ref .
@xcite to model single - spin asymmetries of @xmath1 boson production in realistic rhic conditions .
we put the emphasis on the discussion of the asymmetries @xmath31 and @xmath32 for the distributions in the rapidity @xmath33 and transverse momentum @xmath34 of the observed charged lepton .
we will argue that these directly observed asymmetries provide a viable alternative to the commonly discussed asymmetry @xmath15 .
the paper is organized as follows . in section [ sec : formalismoutline ] , we briefly summarize the resummation formalism used for this study .
a next - to - leading order ( nlo ) calculation is also presented for comparison to the resummation calculation . since good understanding of spin - averaged cross sections
is needed to derive polarized pdfs from the spin asymmetries , section [ sec : unpolarized ] discusses the uncertainties due to the imperfect knowledge of unpolarized parton distributions , as well as the potential for rhic experiments to reduce these uncertainties .
section [ sec : reconstruction - of - spin ] demonstrates that a correct model of multiple parton radiation is important to describe the differential distributions of the final - state leptons .
we then employ the resummation calculation to realistically evaluate the feasibility of the reconstruction of the asymmetry @xmath15 from the observed charged lepton data .
section [ sec : lepton - level - asymmetries ] shows that the spin asymmetries for the distributions of the observed secondary leptons provide an attractive alternative to the asymmetry @xmath15 : they can be measured directly and discriminate efficiently between the different pdf sets .
we also show that the rhic experiments can explore the spin dependence of the nonperturbative contributions from the multiple parton radiation .
the main findings of this paper are recapped in the conclusion .
the primary goal of this paper is to discuss production of @xmath0 bosons , which will be used at rhic to obtain information about the longitudinally polarized parton distribution functions ( pdfs ) in the quark sector . due to the unique feature of the maximal parity violation in the @xmath35 coupling ,
the parton - level cross sections for spin - dependent @xmath1 boson production have pronounced single - spin asymmetries .
this feature gives @xmath1 boson production an edge on parity - conserving processes , in which non - trivial hadronic dynamics can be probed only through the measurement of more complex double - spin asymmetries .
denote the unpolarized distribution of a parton @xmath36 in a nucleon @xmath37 as @xmath38 and the longitudinally polarized distribution of @xmath36 in the nucleon @xmath37 as @xmath39 .
then , @xmath40 where @xmath41 is a helicity - dependent parton distribution function , _ _ i.e. _ _ , a probability of finding a parton @xmath36 with the momentum @xmath42 and helicity @xmath43 in a hadron @xmath37 with the momentum @xmath44 and helicity @xmath45 the pdfs depend on the factorization scale @xmath46 , which in our calculation is assumed to coincide with the qcd renormalization scale . to find the polarized pdfs from an experiment with one longitudinally polarized beam , one measures the unpolarized cross section , @xmath47 as well as the single - spin cross section,@xmath48 here @xmath13 denotes any kinematic variable , such as the rapidity @xmath16 of the @xmath1 boson .
the single - spin asymmetry is defined as @xmath49 in ref .
@xcite , we presented fully differential unpolarized cross sections @xmath50 , single - spin cross sections @xmath51 and double - spin cross sections @xmath52 for production and decay of @xmath53@xmath54 and @xmath3 bosons .
to reliably predict the rate at any point of the momentum phase space , these cross sections include the all - order sum of large soft and collinear logarithms , which dominate at small transverse momentum @xmath26 of the vector boson .
this all - order sum is combined , without double counting , with the finite - order @xmath55 cross section , which dominates at large @xmath26 .
the single - spin resummed cross sections in the narrow width approximation were first obtained in ref .
. the main result of ref .
@xcite is the ( more involved ) derivation of the resummed cross sections that also accounts for the decay of the vector bosons and angular distributions of the final - state leptons .
our result was derived in the @xmath56 factorization scheme @xcite ; therefore , it is fully compatible with the @xmath56 parton distributions . the resummed cross sections from ref .
@xcite are implemented in the numerical resummation program legacy and monte - carlo integration program rhicbos , which were used to produce the results discussed in this paper .
legacy is a c++/fortran program that quickly and accurately generates the resummed cross sections on a grid of points in the kinematical phase space .
rhicbos reads in the cross section grids , performs their multi - dimensional integration , and produces the output that is optimized for rhic specifics . since the finite - order cross section is characterized by a smaller theoretical uncertainty at @xmath57 , rhicbos switches from the resummed cross section ( @xmath58 ) to the finite - order cross section at the point where the resummation cross section becomes smaller than the finite - order cross section at @xmath26 above the position of the maximum in @xmath59 . to compare the resummation calculation to the finite - order approach
, we have also calculated the finite - order @xmath55 cross section ( next - to - leading order , or nlo , cross section ) with the help of a phase space slicing method .
this method introduces a separation transverse momentum scale @xmath60 , below which the integral of @xmath61 over the region @xmath62 is calculated analytically , using the small-@xmath26 approximation in eq .
( 78 ) of ref .
the integral over the region @xmath63 is integrated numerically using the exact @xmath55 cross section . as we will show later , in the nlo approach
, predictions for several important distributions depend strongly on the value of the auxiliary parameter @xmath60 , so that the resummed cross sections must be used in order to model those distributions . some other distributions , such as @xmath64 , are not sensitive to @xmath60 ( up to higher - order corrections ) , so that the resummation and nlo calculations give close results for such distributions .
the electroweak parameters in this paper were obtained in the on - shell scheme according to the procedure described in section ii.b of ref .
@xcite for the fermi constant @xmath65 , @xmath1 boson mass @xmath66 gev , and @xmath3 boson mass @xmath67 gev .
the width @xmath68 of the @xmath1 boson was evaluated as @xmath69 .
the finite - order and asymptotic cross sections were calculated for the qcd factorization scale @xmath70 . the resummed cross section ( see eqs . ( 79 ) and ( 80 ) in ref .
@xcite ) was calculated for the canonical values of the scale parameters , @xmath71 , where @xmath72 is the euler constant , and @xmath73 . unless stated otherwise , the nonperturbative sudakov factor in the resummed cross sections was evaluated by using the recent gaussian parametrization @xcite derived from the global analysis of transverse momentum distributions in unpolarized vector boson production .
the projected statistical errors @xmath74 in the measurement of the single - spin asymmetries were estimated according to eq .
( 13 ) in ref .
@xcite : @xmath75 where @xmath76 is the number of the events for the given luminosity @xmath77 and cross section @xmath78 . unless stated otherwise , the assumed integrated luminosity is @xmath79 , and polarization @xmath80 of the beam @xmath37 is @xmath81 .
.[table : xsect ] the unpolarized cross sections ( in pb ) and numbers of events @xmath82 for the processes @xmath83 and @xmath84 at rhic . here
@xmath85 and @xmath86 are leptons of one lepton generation , and the decay branching ratios were calculated to be @xmath87 @xmath88 .
the numbers in parentheses show the relative uncertainties .
also shown are the born - level momentum fractions @xmath89 corresponding to @xmath90 ( or @xmath91 ) .
the uncertainties @xmath92 in the cross sections @xmath78 are due to the uncertainties in the parton distribution functions estimated according to the cteq6 analysis @xcite .
the uncertainties in @xmath82 are purely statistical . [ cols="^,^,^,^ " , ] to extract the single - spin cross section @xmath93 from the single - spin asymmetry @xmath12 , the spin - averaged cross sections @xmath50 must be known well . obviously , the uncertainty in the knowledge of @xmath50 must be small enough as compared to the targeted uncertainty in the measurement of @xmath93 . therefore , an interesting question is : how well do we currently know the unpolarized cross sections at rhic energies , and how important is their measurement in the upcoming rhic experiments ?
the goal of this section is to argue that at present the uncertainties in the unpolarized cross sections at rhic are sizeable ; to reliably predict the unpolarized rate , these uncertainties must be reduced by dedicated rhic measurements .
table [ table : xsect ] lists the unpolarized cross sections and event rates for the processes @xmath83 and @xmath84 at rhic for various values of center - of - mass energies @xmath94 and integrated luminosities @xmath77 .
the cross sections were derived at nlo using the updated cteq6 parton distribution functions @xcite .
the cross sections @xmath78 are calculated using the best - fit pdf parametrization cteq6 m . the uncertainty @xmath92 in the cross section @xmath78 is generated by the uncertainty in the unpolarized parton luminosities , which are themselves phenomenological functions known with a finite precision .
this uncertainty is evaluated with the help of the hessian matrix analysis employed in ref .
specifically , @xmath92 corresponds to the maximal variation of @xmath78 for all possible pdf parametrizations lying in the acceptable range of @xmath95 . according to table [ table : xsect ] ,
the relative pdf uncertainties are of order @xmath96 at @xmath97 gev and @xmath98 at @xmath6 gev .
these are the errors that exist prior to the measurements at rhic .
the above pdf uncertainties are large because @xmath1 and @xmath99 boson production at the @xmath5 collider rhic probes the sea quark distributions ( primarily @xmath100 and @xmath101 ) at @xmath102 .
for instance , the major part of @xmath1 boson events occurs in the central - rapidity region @xmath103 and corresponds to the born - level momentum fractions @xmath89 close to @xmath104 and @xmath105 at @xmath97 gev and @xmath6 gev , respectively . at such @xmath106 the sea quark pdfs
are not as constrained by the existing ( mostly dis ) data , which better determine the valence - dominated distributions @xmath107 and @xmath108 therefore , rhic provides the information that is complementary to the dis experiments at hera and collider experiments at the @xmath18 collider tevatron .
this point is illustrated by fig .
[ fig : dsig_unpol ] , which shows the relative pdf uncertainty @xmath109 in @xmath0 and @xmath3 boson production at various existing colliders . to compare the potential of @xmath5 and @xmath18 colliders of the same center - of - mass energy ,
this figure also shows the pdf uncertainties at fictitious @xmath18 colliders with @xmath97 and @xmath110 gev . according to the figure
, the pdf uncertainties at rhic substantially exceed the uncertainties at the tevatron and lhc .
they are also up to 2.5 times larger than the pdf uncertainties at the fictitious @xmath18 colliders of rhic energies .
the pdf uncertainty is the largest in @xmath9 boson production at rhic , which is sensitive primarily to the @xmath111 combination of the parton distributions . on the other hand , for the planned integrated luminosities
, rhic will measure the spin - averaged rates with better accuracy , particularly in @xmath8 boson production . at @xmath6 gev and @xmath79 , the @xmath112 relative statistical errors ( @xmath113 ) in the rate
are @xmath114 @xmath115 , and @xmath116 in @xmath8 , @xmath9 , and @xmath3 boson production , respectively . at @xmath97 gev and @xmath117 ,
the respective errors are @xmath118 , @xmath119 and @xmath120 .
we see that the projected statistical errors in @xmath1 boson production are smaller than the current pdf errors .
therefore , we conclude that the measurement of the spin - averaged cross sections at rhic will drastically reduce the uncertainty in the unpolarized parton distributions in the probed region of @xmath106 .
we emphasize that the measurement of the spin - averaged cross sections should be an important part of the rhic physics program , since it complements information about the unpolarized pdfs from the other colliders and reduces systematic uncertainties in the measured polarized cross sections .
to extract useful information from the raw data in @xmath1 boson production , it is necessary to unfold the kinematical acceptance of the signal events , which can only be reliably predicted by a resummation calculation , similar to what has been done for the @xmath1 boson physics at the tevatron @xcite .
the correct lepton - level calculation is even more indispensable given that rhic can not register the transverse energy carried by the decay neutrino . for definiteness ,
let us concentrate on the production of @xmath8 bosons .
of course , all conclusions in this section hold for @xmath9 boson production as well .
[ fig : qt ] shows the transverse momentum distributions of the @xmath8 boson and charged lepton from its decay , as predicted by the resummation package legacy - rhicbos . here and after
, the numerical results correspond to the c.m .
energy @xmath6 gev and integrated luminosity @xmath79 .
it is evident that most of the @xmath1 bosons are produced with small , but non - zero transverse momenta .
such non - zero @xmath26 is acquired through radiation of soft and collinear partons , which can not be approximated by finite - order perturbative calculations . in order to obtain reliable predictions for differential cross sections , dominant logarithmic terms @xmath121 ( where @xmath122 ) associated with such radiation should be summed through all orders of the perturbative series .
furthermore , consider the distribution @xmath123 in fig .
[ fig : qt](b ) .
if the transverse momentum @xmath26 and width @xmath68 of the @xmath1 boson were negligible ( as in the born - level calculation ) , @xmath123 would show a singular jacobian peak exactly at @xmath124 .
the smearing of the jacobian peak in fig .
[ fig : qt](b ) is due to the multiple soft gluon radiation and non - zero width of the @xmath1 boson .
+ ( a)(b ) the example of @xmath123 illustrates that multiple parton radiation and heavy boson decay may drastically alter the born - level predictions .
the resummation effects are important if the distribution is sensitive to @xmath26 directly ( as @xmath123 ) , or if the phase space is constrained in a region with slower convergence of perturbation theory ( as @xmath125 at large @xmath126 ) . in the remainder of this section , we utilize this accurate calculation to evaluate feasibility of the measurement of the boson - level asymmetry @xmath15 , which is commonly discussed in literature as the most convenient observable to probe polarized sea quark distributions . many aspects of our discussion are based on the analysis of two - variable distributions @xmath127 in the transverse momentum and rapidity of the observed charged lepton .
these distributions for the unpolarized and single - spin @xmath8 and @xmath9 boson production are shown in fig .
[ fig : dsigma / dptldyl ] .
consider the born - level expression for the following asymmetry in @xmath8 boson production induced by the up quark and down antiquark via @xmath128:@xmath129 here @xmath130 and @xmath131 is the polar angle of the charged lepton in the rest frame of the @xmath1 boson , with the @xmath132 axis pointing in the moving direction of the polarized proton beam .
the single - spin asymmetry @xmath15 in the rapidity of the @xmath1 boson can be obtained from eq .
( [ alywtheta ] ) by integrating @xmath133 out , i.e. , integrating over the moving direction of @xmath134 and @xmath86 : @xmath135 the resulting expression for @xmath15 is commonly discussed in literature as an example of a simple observable , which is straightforwardly sensitive to @xmath136 and @xmath137 similarly , the born - level asymmetry @xmath15 in @xmath9 boson production is straightforwardly sensitive to @xmath138 and @xmath139 .
it has been also argued in ref .
@xcite that @xmath55 corrections cancel to a good degree between the numerator and denominator of @xmath140 unfortunately , as we will now demonstrate , the measurement of @xmath15 is obstructed by the effects of @xmath1 boson decay and the limited detector acceptance , which seriously complicate both the reconstruction of @xmath16 and determination of @xmath15 from the observed data .
let us first discuss the determination of the rapidity of the @xmath1 boson .
given that neither of the two rhic detectors measures the missing energy , the four - momentum of the @xmath1 boson can not be deduced from the momenta of its decay products .
therefore the rapidity of the @xmath1 boson can not be measured directly . despite the impossibility to measure @xmath16 in general ,
it has been proposed @xcite to statistically _ choose _ the correct @xmath16 _ _ in a certain kinematical region _ _ based on several assumptions about the qcd dynamics of the process .
let us outline the main idea of this method .
denote the rapidity and transverse momentum of the lepton in the lab frame as @xmath33 and @xmath34 , respectively .
similarly , denote the corresponding variables in the rest frame of the @xmath1 boson as @xmath141 and @xmath142 , respectively . in the naive born approximation ,
the @xmath1 bosons have zero width ( @xmath143 ) and zero transverse momentum ( @xmath144 ) .
therefore , the lepton variables in the different reference frames are related as@xmath145{c } = \\
\stackrel{\uparrow } { q_{t}=0}\end{array } p_{t\ell } ^{\prime } , \label{ptl}\end{aligned}\ ] ] where @xmath141 and @xmath142 are the functions of the polar angle @xmath131 of the lepton in the @xmath1 boson rest frame:@xmath146{c } \begin{array}[t]{c } = \\
\stackrel{\uparrow } { q = m_{w}}\end{array } \end{array } \frac{m_{w}}{2}\sin \theta ^{*}.\label{ptlprime}\end{aligned}\ ] ] the comments below the equations show the assumptions involved in obtaining the rightmost expressions in eqs .
( [ ptl ] ) and ( [ ptlprime ] ) . + ( a)(b ) for a given set of @xmath33 and @xmath34 , eqs .
( [ yl])-([ptlprime ] ) imply two solutions @xmath147 and @xmath148 ( corresponding to @xmath131 and @xmath149 ) for @xmath16 .
one of them can be correctly chosen if the magnitude of the rapidity @xmath33 of the charged lepton is large ( i.e. , if the lepton is observed in the forward or backward rapidity region ) . in these special regions , one of the two solutions for @xmath16
can always be discarded , because it takes a value outside of the allowed rapidity range of the collider .
in other words , the chosen solution is the one with the smaller magnitude .
however , in reality , @xmath26 never vanishes , as predicted by the resummation calculation ( see fig . [
fig : qt](a ) ) .
similarly , @xmath68 is not zero .
therefore , the above approximation never holds exactly , and the size of the error is determined by the difference between the born - level and exact dynamics . by comparing the leading order , nlo , and resummation calculations , we have found that the approximation holds reasonably well when @xmath150 , @xmath151 , and @xmath33 is large .
this point is illustrated in fig .
[ fig : butterflies ] , which shows the statistically averaged differences @xmath152 between the exact @xmath16 and each of the two solutions @xmath153 in the two - dimensional plane of @xmath33 and @xmath34 , predicted by the resummation calculation .
it can be seen that the difference is small for one of the solutions in the respective region @xmath154 or @xmath155 , and @xmath156 gev . at the same time , the difference is unacceptably large for the other solution .
correspondingly , in each special region the distribution in the correct solution @xmath147 or @xmath148 closely reproduces the distribution in true @xmath16 .
for example , fig .
[ fig : y_w_app ] shows that the distributions @xmath157 agrees well with the distribution @xmath158 in the region @xmath159 and @xmath160 . except for the special regions in fig .
[ fig : butterflies ] , the approximate solutions @xmath153 do not agree well with the exact @xmath16 .
an important feature to note is that in the special regions the event rate is small . as shown in fig .
[ fig : dsigma / dptldyl ] , most of the event rate comes from @xmath161 and @xmath162 .
furthermore , the magnitude of the error in @xmath16 depends on the shape of the cross section , which , of course , is unknown _ a priori _ in the polarized case .
let us now turn to another problem in the measurement of @xmath15 : distortions in the deduced shape of @xmath15 due the experimental cuts . to obtain @xmath15
, the experimental cross sections have to be integrated over all possible directions of the lepton motion , i.e. , over @xmath133 in the interval @xmath164 ( see eq .
( [ alyw ] ) for an example of such integration in the born - level analysis ) .
unfortunately , the acceptance of rhic detectors does not permit such integration , so that the resulting asymmetry has the residual dependence on the acceptance range in @xmath165 and does not agree with the true asymmetry @xmath15 .
indeed , @xmath133 is related to the rapidity of the lepton @xmath33 via eqs .
( [ yl ] ) and ( [ ylprime]):@xmath166 therefore , the restrictions on the range of @xmath33 reduce the range of integration over @xmath167 specifically , the muon detectors are able to detect a muon if the muon rapidity @xmath168 and azimuthal angle @xmath169 satisfy @xmath170 and @xmath171
. the electromagnetic calorimeter can register electrons with @xmath172 and @xmath173 . the detector registers electrons in the barrel electromagnetic calorimeter that covers @xmath174 and @xmath175 . in addition , a selection cut @xmath176 gev is expected to be imposed on the charged leptons to suppress background contributions , which also constrains @xmath177 through eqs .
( [ ptl ] ) and ( [ ptlprime ] ) .
the effect of the above cuts is demonstrated in fig .
[ fig : alyvscuts ] , which shows the single - spin asymmetry in @xmath8 boson production calculated without constraints on @xmath33 and @xmath34 ( solid line ) , as well as with the constraints @xmath178 @xmath179 ( circles ) and @xmath180 @xmath181 gev ( boxes ) .
the results shown here are derived using the resummation calculation .
, the resummation and nlo calculations give close predictions ( within one estimated statistical error @xmath74 ) , even though these calculations predict quite different shapes for the unpolarized and polarized differential cross sections . ]
according to the figure , there is a substantial difference between the asymmetries calculated with and without experimental cuts .
this difference arises due to the different dependence of the unpolarized and polarized cross sections on angular distributions of the charged leptons , which affects the asymmetry because @xmath133 is not integrated out completely .
the findings in this section can be summarized as follows .
the extraction of @xmath15 from rhic data is not an easy task , because ( a ) the direct measurement of @xmath16 is impossible , and ( b ) the reconstructed @xmath15 is distorted by the experimental acceptance cuts .
in particular , the two - fold ambiguity in the determination of @xmath16 is unavoidable at @xmath162 and @xmath182 i.e. , in the region with the largest cross section . to take the full advantage of the large event rate at rhic , we ought to seek an alternative to @xmath15 , which is free of the complications discussed above . in the next section
, we turn our attention to the spin asymmetries of the lepton - level cross sections @xmath125 and @xmath123 , which may serve as such alternative observables .
based on the resummation calculation , we analyze the sensitivity of the lepton - level asymmetries to the choice of the experimental cuts .
we also present theoretical predictions for various pdf sets .
let us first separately discuss the unpolarized and polarized cross sections . to study the detector effects on the shape of various distributions , we will compare the cross sections for the kinematical acceptance of and detectors , as well as cross sections without any kinematical cuts .
we also compare the resummation calculation to the nlo calculation , with the latter performed in the phase space slicing method for the separation parameter @xmath183 gev .
we concentrate on the distributions @xmath123 , where the differences between the resummed and nlo predictions are the most visible . fig .
[ fig : ptl_unp_wp ] presents the distribution @xmath184 for the positively charged leptons from the decay of @xmath8 bosons .
the shape of the nlo cross sections ( dashed curves ) for @xmath34 around and above @xmath185 ( i.e. , the jacobian peak ) is strongly affected by the arbitrary parameter @xmath60 employed in this calculation .
hence , the only reliable predictions for this part of the phase space are given by the resummed cross sections ( shown as the solid curves ) .
for smaller @xmath34 , the resummation calculation agrees with the nlo calculation , because it is formulated to match the nlo calculation in the region where the multiple gluon emission is not important .
furthermore , without any kinematic cuts , the integrated rate of the resummation calculation agrees with that of the nlo calculation to a percent level . as shown in fig .
[ fig : ptl_unp_wp](b ) , the selection of the leptons in the large rapidity region distorts the shape of the @xmath34 distribution by suppressing the contributions near the jacobian peak . on the other hand , the selection of the central - rapidity region @xmath186 ( cf .
[ fig : ptl_unp_wp](c ) ) suppresses the contributions with @xmath187 gev .
both features can be understood from the shape of the differential distribution @xmath127 given in fig .
[ fig : dsigma / dptldyl](a ) . according to this figure
, the small ( large ) rapidity @xmath126 statistically corresponds to the large ( small ) transverse momentum @xmath34 . as a @xmath5 collider ,
rhic produces different numbers of @xmath9 and @xmath8 bosons because of different parton luminosities .
due to the different rapidity distributions for @xmath8 and @xmath9 bosons , the rapidity cuts have a different effect on @xmath123 in these processes .
[ fig : ptl_unp_wm ] shows the distributions @xmath123 in the process @xmath188 .
we find that the shape of the distributions without cuts shown in fig .
[ fig : ptl_unp_wm](a ) is very close to the shape of the corresponding distribution in fig .
[ fig : ptl_unp_wp](a ) , even though the overall normalizations are obviously different . on the other hand
, the shapes do not coincide in the presence of the cuts , in particular in the case of the selection cuts shown in figs .
[ fig : ptl_unp_wp](b ) and [ fig : ptl_unp_wm](b ) .
the @xmath34 distributions in the single - spin @xmath8 boson production are shown in fig .
[ fig : ptl_pol_wp ] . the cross sections @xmath189 in @xmath8 boson production are negative , because the charged weak coupling is purely left - handed , and the dominant polarized pdf @xmath136 is positive in the rhic range of @xmath106 .
again , we see that the nlo calculation can not reliably predict the distribution of @xmath34 around and above @xmath185 . for the kinematics , cf .
[ fig : ptl_pol_wp](b ) , a sharper peak is developed in the distribution of @xmath34 around @xmath185 as compared to the unpolarized case , cf .
[ fig : ptl_unp_wp](b ) .
the different shape is caused by the different rapidity dependence in the unpolarized and single - spin cases , as shown in figs .
[ fig : dsigma / dptldyl](a ) and [ fig : dsigma / dptldyl](c ) .
finally , the distribution @xmath123 for the charged leptons @xmath190 in the @xmath9 events is shown in fig .
[ fig : ptl_pol_wm ] . for the rapidity distributions
@xmath125 without imposed cuts on @xmath34 , the nlo and resummed cross sections are close to one another , in accordance with the cancellation of soft contributions in sufficiently inclusive observables . however , according to the discussion above , the resummation and nlo predictions for @xmath125 can be very different if cuts on @xmath34 are imposed .
we now turn our attention to the single - spin asymmetry . since the only straightforward signature of the @xmath1 bosons at rhic is the observation of the secondary charged leptons , it is important to understand these asymmetries at the lepton level . in figs .
[ fig : alyl_wp ] and [ fig : alyl_wm ] , we show the lepton - level asymmetries @xmath31 in the @xmath8 and @xmath9 events , respectively , for various cuts on @xmath34 . in these figures , the positive rapidity direction is in the moving direction of the polarized proton beam .
similarly , figs .
[ fig : alptl_wp ] and [ fig : alptl_wm ] show the asymmetries @xmath32 for various cuts on @xmath33 .
all asymmetries are derived using the gehrmann - stirling pdf sets a and b @xcite and grsv-2000 valence - like pdf set @xcite in the resummation calculation .
we found by experimenting with other available pdf sets in refs .
@xcite that the shown graphs characterize well the possible span in the asymmetry due to the different polarized pdf parametrizations .
the projected statistical errors are estimated according to eq .
( [ deltaal ] ) for the bin sizes shown in the figure .
according to the figures , the lepton - level asymmetries can be accurately measured for both @xmath8 and @xmath9 bosons .
these directly observed asymmetries can efficiently discriminate between different pdf sets ; hence , they provide a viable alternative to the less accessible asymmetry @xmath15 . for both @xmath8and @xmath9 boson samples ( cf
. figs .
[ fig : alyl_wp ] and [ fig : alyl_wm ] ) , the power to discriminate between different pdf sets is larger in the region of the positive rapidity @xmath33 , especially for the @xmath9 boson data . it is interesting to observe that , in different ranges of @xmath191 the largest variations due to the choice of the pdfs appear in different ranges of @xmath34 . in @xmath8 boson production , the largest variation in the asymmetries @xmath32
is concentrated at @xmath192 gev in the backward rapidity region @xmath193 ( fig .
[ fig : alptl_wp](a ) and ( b ) ) and at @xmath194 gev in the forward rapidity region @xmath195 ( fig .
[ fig : alptl_wp](c ) and [ fig : alptl_wp](d ) ) . in @xmath9 boson production the largest variation
is observed for very forward leptons , @xmath196 ( fig .
[ fig : alptl_wm](d ) ) .
if an additional condition @xmath181 gev is imposed , other regions of @xmath33 ( cf . fig .
[ fig : alptl_wm](a)-[fig : alptl_wm](c ) ) show a smaller sensitivity to different pdf sets . during the first years of rhic spin program
, it may not be possible to distinguish between the positively and negatively charged leptons from the @xmath1 boson decay , particularly before the solenoid field of the detector is fully operational .
it is therefore interesting to examine the possibility of measuring the single - spin asymmetries in the combined sample of @xmath197 and @xmath9 bosons , without distinguishing between charged leptons and anti - leptons .
such asymmetries are shown in figs .
[ fig : alyl_wpwm ] and [ fig : alptl_wpwm ] .
since there are about three @xmath8 bosons produced per each @xmath9 boson , these asymmetries are close to the @xmath8 boson asymmetries shown in figs .
[ fig : alyl_wp ] and [ fig : alyl_wm ] .
therefore , these asymmetries will be able to distinguish between different pdf sets for a sufficiently high number of events . and
[ fig : alptl_wpwm ] ( evaluated for @xmath79 ) will have to be adjusted to agree with the actual integrated luminosity at the time of the measurement .
] however , they can not completely replace the single - spin asymmetries for independent @xmath8 and @xmath9 boson samples , which have better discriminating power and allow us to distinguish between contributions of different parton flavors .
as was discussed above , the transverse momentum distributions for vector bosons are affected by the soft and collinear radiation , which can be described only by means of all - order resummation .
in addition , the distributions at very small @xmath26 are sensitive to nonperturbative contributions characterized by large impact parameters @xmath198 . as a result
, the shape of the lepton - level distribution @xmath123 around its peak at about @xmath124 is affected by both perturbative and nonperturbative qcd radiation , in addition to the nonzero width of the @xmath1 boson .
while the perturbative soft and collinear contributions can be calculated order - by - order in the resummation formalism , the nonperturbative contributions ( associated with the nonperturbative transverse motion of partons inside the proton ) can only be approximated by phenomenological parametrizations .
in particular , based on the factorization properties of the resummed cross sections , one can hypothesize a simple relationship between the nonperturbative sudakov factors in the unpolarized , single - spin , and double - spin vector boson production:@xmath199 furthermore , @xmath200 is usually assumed to be independent of the type of the electroweak hard probe , so that it is the same in @xmath53 @xmath201 and @xmath99 boson production .
if the relationship ( [ snpcrossing ] ) for the spin dependence of @xmath200 indeed holds in nature , the nonperturbative contributions in single - spin @xmath1 boson production can be determined unambiguously once the parametrizations of @xmath200 on the right - hand side of eq .
( [ snpcrossing ] ) are available , e.g. , from the global analysis of the distributions @xmath59 in the unpolarized and double - spin @xmath202 or @xmath3 production .
it is well known from the unpolarized studies ( see , e.g. , refs .
@xcite ) that the importance of the nonperturbative contributions in the massive boson production at high - energy colliders is strongly reduced as compared to drell - yan pair production in fixed - target experiments . nonetheless , this dependence is not entirely negligible at small transverse momenta of @xmath1 bosons , roughly for @xmath203 gev or less .
to give an idea about the typical size of the nonperturbative contributions , fig .
[ fig : ptl ] shows the event rate for the single - spin cross section @xmath189 at @xmath79 , integrated over the acceptance range @xmath204 .
this rate was calculated using two parameterizations @xcite of the nonperturbative part , which were found from the analysis of unpolarized vector boson production .
correspondingly , the resummed cross sections were calculated with the assumption that @xmath205 .
it can be seen that the sensitivity to the nonperturbative input is small , but , nonetheless , visible near the jacobian peak . for comparison
, we also included the nlo cross section calculated using the phase space slicing method .
the nlo curve substantially deviates from the resummation curves , and , moreover , its shape can be drastically modified by varying the phase space slicing parameter @xmath60 .
in contrast , practically all features of the resummation curve are determined unambiguously by the perturbative contribution , and the remaining small dependence on the nonperturbative contributions is pinned down by utilizing a phenomenological parameterization discussed above .
needless to say , rhic should explore the spin properties of nonperturbative contributions by studying distributions @xmath59 in the double - spin @xmath202 and @xmath3 production and @xmath123 in single - spin and double - spin @xmath1 boson production .
due to the maximal parity violation in the weak current coupling , the longitudinal single - spin asymmetry of @xmath1 boson production is strongly sensitive to the polarized parton distributions . at the center - of - mass energy of 500 gev , @xmath1 boson production events at rhic
can clearly discriminate between various polarized quark luminosities . as a @xmath5 collider , rhic probes the sea quark pdfs in the region of relatively high @xmath206 where such pdfs are not adequately constrained even in the unpolarized case .
therefore , rhic experiments also have a good potential to reduce the uncertainties in the unpolarized sea quark pdfs in a dedicated study of spin - averaged cross sections . since and detectors do not determine the missing transverse energy associated with the neutrino from @xmath1 boson decay , and because the charged leptons from the decay can only be observed in a part of the @xmath207 solid angle , the reconstruction of the asymmetries at the level of @xmath1 bosons is not easy to realize . as an alternative to the asymmetry @xmath15 for the @xmath1 boson rapidity distribution @xmath158 , we advocate the measurement of the asymmetries @xmath208 ) and @xmath32 for the distributions @xmath125 and @xmath123 in the rapidity @xmath33 and transverse momentum @xmath34 of the secondary charged lepton .
these lepton - level asymmetries can be directly observed at rhic detectors , they are sizeable , and they can be used to determine the polarized pdfs with high precision . to reliably predict the distributions of the charged leptons in all available phase space , a theoretical calculation is needed for all - order summation of large logarithmic contributions corresponding to @xmath1 bosons with a small transverse momentum . at the lepton level , the largest effect of the logarithm resummation
is observed in the distributions @xmath123 and @xmath189 in the proximity of the jacobian peak ( @xmath209 ) .
the importance of the resummation can be verified by observing a large difference between the resummed and nlo cross sections near the jacobian peak , which happens only when the effect of soft gluon radiation is important . in this paper , we apply the state - of - art resummation calculation @xcite for spin - dependent vector boson production to predict reliably the distributions of the observed leptons from @xmath1 boson decay .
in addition to the spin - averaged and single - spin cross sections , we presented a detailed study of single - spin asymmetries for the decay lepton in the presence of kinematical cuts imposed by and detectors .
furthermore , we propose to study the distribution @xmath189 at @xmath209 to learn about the spin ( in)dependence of nonperturbative contributions in the css resummation formalism .
the study of such contributions for different beam polarizations and bosons of different types ( @xmath53@xmath0 , @xmath3 ) will shed light on basic properties of qcd factorization and provide important clues about the nature of the intrinsic transverse motion of partons inside the proton .
authors would like to thank c. balazs , d. boer , g. bunce , m. grosse perdekamp , s. gupta , j. kiryluk , n. saito , m. stratmann , w. vogelsang , and members of the cteq collaboration for helpful discussions .
we are grateful to f. olness for his comments on the manuscript .
we thank the organizers of rhic spin workshops , where preliminary results of this work were presented . the work of p. m. n. has been supported by the u.s .
department of energy and lightner - sams foundation .
the research of c. p. y. has been supported by the national science foundation under grant phy-0100677 . | in order to measure spin - dependent parton distributions in production of @xmath0 bosons using the non - hermetic detectors of the relativistic heavy ion collider , an accurate model for distributions of charged leptons from the @xmath1 boson decay is needed .
we discuss the predictions for production and decay of the @xmath1 bosons based on a calculation for resummation of large logarithmic contributions originating from multiple soft gluon radiation .
we compare the predictions for the spin - dependent resummed cross sections with the corresponding next - to - leading order results .
we show that the lepton - level asymmetries can be reliably predicted by the resummation calculation and directly compared to the experimental data . a program for the numerical analysis of such cross sections in @xmath2 and @xmath3 boson production is also presented . |
on note @xmath1a , b[$ ] o @xmath2 et @xmath3 $ ] ladhrence de @xmath4 dans @xmath5 .
soit @xmath6 un corps et @xmath7 . on se donne un espace probabilis @xmath8 sur lequel existent une
famille croissante de tribus @xmath9 et une famille dcroissante de tribus @xmath10 . on note @xmath11 lensemble des processus @xmath12 dfinis sur @xmath13 , valeurs dans @xmath14 et tels que : @xmath12 soit @xmath15 et @xmath16 adapt , pour tout @xmath17 @xmath18 , lapplication @xmath19 de @xmath20 dans @xmath21 est continue , pour tout @xmath22 les quantits @xmath23 $ ] , et @xmath24 $ ] , existent dans @xmath21 ,
et enfin les applications @xmath25 et @xmath26 sont continues de @xmath4 dans @xmath21 .
+ le complt de @xmath27 pour la norme @xmath28 , est encore not @xmath11 , et simplement @xmath29 quand @xmath30 . on note @xmath31 la
drive stochastique introduite dans @xcite et dfinie par @xmath32 . on dira
quun lagrangien @xmath33 est admissible si la fonction @xmath34 est dfinie sur @xmath35 , @xmath36 en @xmath37 et holomorphe en @xmath38 , et est relle quand @xmath38 est rel .
@xmath33 est dit naturel sil scrit @xmath39 o @xmath40 est une forme quadratique sur @xmath41 et @xmath42 un potentiel de classe @xmath43 sur @xmath44 .
soit alors la fonctionnelle associe @xmath33 dfinie par @xmath45 , @xmath46 $ ] avec @xmath47 des processus @xmath33-adapt @xmath48 .
soit @xmath49 un sous espace de @xmath29 . on appelle
@xmath50variation dun processus @xmath51 , un processus de la forme @xmath52 o @xmath53 .
soit @xmath54 le sous - espace vectoriel de @xmath49 dfini par @xmath55 , et @xmath56 le sous - espace vectoriel @xmath57 .
posons alors la si @xmath33 est un lagrangien admissible et @xmath58 la fonctionnelle associe , @xmath58 est dite @xmath49-diffrentiable en un processus @xmath59 si pour tout @xmath60 , @xmath61 , o @xmath62 est une fonctionnelle linaire en @xmath63 et @xmath64 .
de plus @xmath12 est dit @xmath50stationnaire si pour tout @xmath60 , @xmath65 .
on considre le cas @xmath66 .
la @xmath56-diffrentielle de @xmath58 est donne par : soit @xmath33 un lagrangien admissible do nt toutes les diffrentielles secondes sont bornes .
posons @xmath67 $ ] .
alors @xmath68 et la fonctionnelle @xmath58 associe @xmath33 est @xmath69-diffrentiable en tout processus @xmath70 , et pour tout @xmath71 , sa diffrentielle scrit : + @xmath72 + g(z,\partial_v l)(b)-g(z,\partial_v l)(a)$ ] .
* dmonstration . *
laide du dveloppement de taylor de @xmath33 , on a @xmath73 o @xmath74 .
daprs lingalit de cauchy - schwarz , @xmath75 $ ] , @xmath76 $ ] et @xmath77 $ ] sont des @xmath78 car @xmath79 .
de plus @xmath80 , @xmath81 et @xmath82 sont bornes et @xmath70 . on en dduit que @xmath68 et @xmath83+o(\|z\|)$ ] .
on conclut en utilisant ( [ formule2 ] ) et le fait que @xmath84 .
@xmath85 rappelons les lemmes suivants respectivement dmontrs dans @xcite p.42 et @xcite p.70 , le premier gnralisant la `` loi produit '' donne par nelson dans @xcite p.@xmath86 , le deuxime tant lanalogue stochastique du lemme classique des `` fonctions plateaux '' ( _ cf _ @xcite p.@xmath87 ) : [ formule2 ] soit @xmath88 .
alors @xmath89 = \frac{d}{dt}e[x_t\cdot y_t]$ ] .
soit @xmath90 .
si pour tout @xmath84 @xmath91du=0 $ ] , alors @xmath92 est constant . des trois lemmes on dduit le une condition ncessaire et suffisante pour quun processus @xmath93 soit un processus @xmath94stationnaire pour @xmath58
est quil vrifie lquation deuler - langrange stochastique @xmath95 : @xmath96 sur @xmath20 .
le calcul des variations stochastiques est dvelopp ici indpendemment de la procdure de plongement stochastique dfinie dans @xcite .
il devient cohrent avec cette dernire pour un choix adquat de lespace de variation , en ce sens que : le diagramme suivant commute : @xmath97_{\bigstar } \ar[r]^{\cal{s } } & l(x_t,\mathcal{d}x_t ) \ar[d]^ { \maltese } \\ & ( el ) \ar[r]_{\cal{s } } & ( els ) } \end{aligned}\ ] ] o @xmath98 symbolise lutilisation du principe de moindre action classique partir de la fonctionnelle + @xmath99 , @xmath100 celle du @xmath101-principe de moindre action stochastique partir de @xmath102 $ ] et @xmath103 la procdure de plongement stochastique dfinie dans @xcite .
dans @xcite chap.@xmath104 , on traite le cas @xmath105 et on montre quune condition ncessaire et suffisante pour quun processus @xmath93 soit un processus @xmath106stationaire pour @xmath58
est quil vrifie lquation : @xmath107 sur @xmath20 .
ce cas ne nous permet pas dobtenir un lemme de cohrence mais un thorme de noether stochastique .
les symtries apparaissant dans certains systmes lagrangiens induisent lexistence dintgrales premires du mouvement ( _ cf _ @xcite p.@xmath108 ) .
une question naturelle est alors dtudier la persistance de ces objets fondamentaux sur le systme stochastis par la procdure de plongement .
introduisons lensemble @xmath109 des processus dfinis sur @xmath13 et lespace @xmath110 do nt la dfinition est donne dans @xcite p.@xmath111 . on montre suivant @xcite que lon peut calculer les drives @xmath112 et @xmath113 sur des lments de cet espace ( _ cf _
@xcite p.@xmath114 ) .
on rappelle le soit @xmath121 un diffomorphisme .
la suspension stochastique de @xmath122 est lapplication @xmath123 dfinie par @xmath124 .
dans la suite on notera indiffremment le diffomorphisme et sa suspension .
+ de plus un groupe un paramtre de transformations @xmath125 , @xmath126 , o @xmath127 , est appel un groupe @xmath122-suspendu agissant sur @xmath128 sil existe un groupe un paramtre
de diffomorphisme @xmath129 , @xmath126 , tel que pour tout @xmath126 , @xmath130 soit une suspension stochastique de @xmath131 , et pour tout @xmath132 , @xmath133 .
un groupe un paramtre de diffomorphismes est dit admissible si @xmath134 est un groupe un paramtre de @xmath135-diffeomorphismes sur @xmath44 tel que @xmath136 et tel
que la formule ( [ deriv_fonc ] ) reste vrai pour toute fonction @xmath131 du groupe .
la dernire condition peut paratre trs restrictive , mais elle est vrifie pour les groupes un paramtre de diffomorphismes affines de @xmath44 , ce qui est important dans le cas classique ( @xcite p.@xmath137 ) .
[ tecnicnoether ] soit @xmath138 une suspension stochastique dun groupe admissible un paramtre de diffomorphismes .
alors pour tout @xmath139 , et pour tout @xmath140 lapplication @xmath141 est de classe @xmath36 @xmath142 et @xmath143=\mathcal{d } \left [ \partial_s \phi_s(x)\right]$ ] @xmath142 .
soit @xmath144 ne suspension stochastique dun groupe admissible un paramtre de diffomorphismes et @xmath145 .
la fonctionnelle @xmath33 est invariante sous @xmath146 si pour tout @xmath147 et @xmath51 , @xmath148 .
[ noether ] soit @xmath58 la fonctionnelle dfinie sur @xmath149 par @xmath150 $ ] , o @xmath33 est un lagrangien admissible invariant sous le groupe admissible un paramtre de diffomorphisme @xmath138 .
soit @xmath151 un point @xmath29-stationnaire de @xmath58 . on pose @xmath152 .
alors @xmath153=0 $ ] . *
dmonstration . * on pose @xmath154 .
comme @xmath33 est invariant
sous @xmath155 , on a @xmath156 comme pour tout @xmath17 et tout @xmath157 , @xmath158 et @xmath159 , on obtient @xmath160 . en utilisant le lemme ( [ tecnicnoether ] ) , cette quation est quivalente @xmath161 .
comme @xmath162 est un @xmath106point stationaire de @xmath58 , on a @xmath163 $ ] . on en dduit alors @xmath164 \cdot { \partial y_t \over
\partial s}(0 ) + \partial_v l(v_t(0 ) ) \cdot { \cal d } \left ( { \partial y_t \over \partial s}(0)\right ) = 0 \quad ( p.s.)$ ] .
do @xmath165 \cdot { \partial y
\over \partial s}(0 ) + \partial_v l(v_t(0 ) ) \cdot { \cal d } \left ( { \partial y_t \over \partial s}(0)\right ) \right ] = 0 $ ] . + avec le lemme ( [ formule2 ] ) , il vient @xmath166 = 0 $ ] .
@xmath85 soit @xmath33 un lagrangien admissible .
une fonctionnelle @xmath167 est une intgrale premire pour lquation d euler - lagrange stochastise associe @xmath33 si @xmath168 = 0,$ ] pour tout @xmath12 satisfaisant une quation deuler - lagrange stochastique
. 7 arnold v.i .
, _ mathematical methods of classical mechanics _ , 2d edition , springer , 1989 .
cresson j. , darses s. , stochastic embedding of dynamical systems , arxiv : math.pr/0509713 , 112.p , 2005 .
cresson j. , darses s. , plongement stochastique des systmes lagangiens , 4.p , 2005 .
nelson e. , _ dynamical theories of brownian motion _ , second edition , princeton , 2001 .
thieullen m. , second order stochastic differential equations and non - gaussian reciprocal diffusions , proba .
theory and rel .
fields 97 , 231 - 257 ( 1993 ) .
yasue k , stochastic calculus of variations , journal of functional analysis 41 , 327 - 340 ( 1981 ) . | la procdure de plongement stochastique , dfinie dans @xcite , permet dassocier lquation deuler - lagrange classique ( el ) une quation deuler - lagrange stochastique ( els ) .
cette dernire est - elle sous - tendue par un principe de moindre action gnralis ?
pour aborder cette question , nous dveloppons un calcul des variations stochastique , initi par yasue @xcite .
on donne un analogue stochastique naturel @xmath0 de la fonctionnelle lagrangienne daction .
on introduit une notion de stationnarit pour laquelle les solutions de ( els ) sont les points stationnaires de @xmath0 .
la notion de stationnarit ainsi dfinie rend cohrent le calcul des variations stochastique vis - - vis de la procdure de plongement stochastique .
enfin , nous dmontrons un thorme de noether stochastique qui suggre dintroduire une nouvelle notion , celle dintgrale premire stochastique .
_ pour citer cet article : a. nom1 , a. nom2 , c. r. acad .
sci .
paris , ser .
i 340 ( 2005 ) . _ 0.5 * abstract * 0.5*coherence lemma and stochastic noether theorem . * the stochastic embedding procedure defined in @xcite associates a stochastic euler - lagrange equation ( sel ) to the standard euler - lagrange equation ( el ) .
can we derive ( sel ) from a generalized least action principle ? to address this question , we develop a stochastic calculus of variation initiated by yasue @xcite . we give a stochastic analog @xmath0 of the lagrangian action functional .
we introduce a notion of stationarity according to which the solutions of ( sel ) are the stationary points of @xmath0 .
this notion of stationarity brings coherence to stochastic calculus of variation with respect to stochastic embedding . finally , we prove a stochastic noether theorem which introduces an original notion of stochastic first integral . _ to cite this article : a. nom1 , a. nom2 , c. r. acad .
sci .
paris , ser .
i 340 ( 2005 ) . _ probability theory / dynamical systems , |
the instability of an electron gas due to strong correlations at the conditions of dhva effect resulting in diamagnetic phase transition ( dpt ) into inhomogeneous diamagnetic phase ( idp ) with formation of condon domains ( cds ) @xcite , @xcite is intensively studied both theoretically and experimentally @xcite-@xcite .
the realization of intrinsic structure of idp is governed by the competition between long - range dipole - dipole interaction and short - range interaction related to the positive interface energy with typical magnetic length of larmor radius .
the observed idp in ag by splitting of nmr line @xcite was identified as periodic domain structure with alternating in neighboring domains one - component magnetization .
the further experiments on observation of idp by method of @xmath2sr spectroscopy revealed the arise of diamagnetic instability in be , sn , al , pb and in @xcite-@xcite , giving strong support to the idea of formation of cd structure @xcite .
the experimental demonstration of existence of idp in normal metals @xcite,@xcite-@xcite is very convincing and support the basic ideas of the theory of dpt developed in @xcite , @xcite , but the full understanding of the properties of idp is still lacking .
the attempts of the quantitative analysis of the data on the basis of model of plane - parallel band patterns reveal the contradictions @xcite , put questions on correctness of applicability of demagnetizing coefficient for adequate description of shape - dependent properties of idp @xcite , @xcite , @xcite and raised important unanswered questions concerning the type of the diamagnetic ordering , irreversibility of dpt and morphology of the domain patterns .
there are striking similarities between idp in normal metals and other strongly correlated systems which undergo phase transition on temperature and magnetic field with formation of complex macroscopic patterns , e. g. the type - i superconductors and thin magnetic films .
the different technics , including the powder pattern and magneto - optic methods , successfully used for observation of intermediate state of type - i superconductors revealed a very rich structure @xcite which in spite of all it complexity , amazingly reminds the variety of domain structures in thin film @xcite .
recent experimental observation of cds by a set of micro hall probes at the surface of a plate - like sample of ag @xcite demonstrates the formation of complex structure which differs from the expected regular cd patterns separated by plane - parallel dws @xcite and calls for further development of the experimental techniques to study the idp .
the first experimental observation of such an exotic phenomena as diamagnetic hysteresis by complex method including direct measurements by hall probe , standard ac method with different modulation levels , frequencies and magnetic field ramp rates @xcite allows to reconstruct the magnetization reversal in be and confirms the possibility of the first order phase transition .
the giant parametric amplification of non - linear response which proves to be sensitive method for investigation of phase transitions in superconductors was observed in a single crystal of be in quantizing magnetic field @xcite , where the measured output signal amazingly follows the shape of the dpt diagram and shows the asymmetry and shift to the upper edge of the dhva period in ascending magnetic field justifying the possibility of the first order dpt and necessity of further development of the theory of dpt .
motivated by recent data on observation of dpt instability @xcite-@xcite , in the framework of catastrophe theory we investigate the equilibrium set for the system of strongly correlated electron gas at the conditions of dhva effect .
we show that in every period of dhva oscillations the discontinuities of order parameter accompanied dpt is handled by riemann - hugoniot catastrophe implying the standard scenario for the transition , e.g. dpt is of the second order at the center of dhva period , weakly first order in the nearest vicinity of this point and is of the first order at the rest part of the dhva period both on @xmath3 and @xmath4 @xmath5 .
thus , similar to other magnetic systems , e. g. the spin @xcite and metamagnetic ones @xcite , dpt can be realized in a number of different nonuniform phases depending on the shape of the sample . at @xmath6 0 k the fine
, sub - band , structure of landau level can results in dpt of the first order on @xmath4 @xmath5 @xcite , but this effect based on treating the cd instability as an electron topological transition is negligibly small in comparing with the results obtained in the present publication .
we show that the condition for dpt occurrence @xmath71 where @xmath8 is a differential magnetic susceptibility is valid only in the nearest vicinity of the center of dhva period , @xmath90 , where @xmath10 is an increment of magnetic field @xcite .
this condition is violated for @xmath11 and is replaced by the generalized condition on @xmath1 in accordance with the bifurcation set of cusp catastrophe .
we discuss the conditions for realization of the dpt in the sample with taking into account the long - range dipole - dipole interaction and compare the results with available data .
[ t ] as a function of inverse differential magnetic susceptibility @xmath12 at three different values of magnetic field @xmath90 ( a ) , @xmath13/36 ( b ) and @xmath13/3 ( c ) .
the solid ( dash ) lines correspond to the local minimum ( maximum ) of the free energy density @xmath14 .
symmetric supercritical pitchfork bifurcation ( a ) of the system undergoing the second order phase transition at the center of dhva period @xmath15 for @xmath16 ( critical point ) is replaced by the imperfect bifurcation diagram ( b , c ) with increase of @xmath17 ( @xmath18 is a period of dhva oscillations in reduced units ) when the system is a subject for the first order phase transition.,title="fig:",scaledwidth=40.0% ]
equilibrium properties of strongly correlated electron gas at the conditions of dhva effect in one - harmonic approximation can be described by gibbs free energy @xmath19 which is a functional of two conjugated variables @xmath10 and @xmath20 @xcite . here , the small - scale field @xmath21 $ ] is the increment of the large - scale internal magnetic field @xmath22 and applied magnetic field @xmath23 , @xmath24 is oscillating part of reduced magnetization , @xmath25
, @xmath26 is the fundamental frequency of the dhva oscillations corresponding to the extreme cross - section of fermi - surface , and @xmath27 is dhva period .
minimization of the free energy @xmath14 with respect to @xmath24 leads to well - known expression for the magnetization @xcite @xmath28 with the magnetic flux density @xmath29 being a function of increment of internal magnetic field @xmath10 and reduced magnetization @xmath24 .
internal magnetic field @xmath30 depends on applied magnetic field @xmath31 and stray field @xmath32 , which take into account the long - range dipole - dipole interaction dependent on the shape of a sample and has to be found by solving the maxwell equations in magnetostatic approximation by a self - consistent procedure . as it was suggested by shoenberg @xcite , the magnetic flux dependence of oscillating part of magnetization @xmath24 eq .
( [ eq : magnetization ] ) at curtain conditions , when the differential susceptibility @xmath331 , results in arising instability of uniform phase ( shoenberg effect ) . at this condition both the magnetization and magnetic induction become a multi - valued function of internal magnetic field @xmath10 .
therefore , depending on long - range dipole - dipole interaction at high magnetic field and low temperature the system of strongly correlated electron gas can undergo the dpt with formation of idp which corresponds to a global minimum of the free energy of the system .
due to the governing role of the long - range dipole - dipole interactions , the morphology of the dpt depends crucially on the shape of the sample which , in general , can not be account for by the straightforward use of conception of demagnetizing coefficient @xcite .
the magnetization @xmath24 eq .
( [ eq : magnetization ] ) as a function of inverse reduced amplitude of the dhva oscillations @xmath12 at three values of the increment of magnetic field @xmath90 , @xmath343 and 2@xmath343 is plotted in fig .
[ magnetization ] , which illustrates the typical bifurcation behavior of the system , e. g. the existence of two stable states for magnetization at some values of the parameters @xmath351 and @xmath10 , and the possibility of discontinues change of the order parameter through the disappearance of stable states when @xmath36 .
the discontinuities of the function @xmath37 where parameter @xmath1 controls the amount of ordering , or the value of order parameter , and parameter @xmath10 breaks the @xmath38 symmetry of order parameter , can be handled in the framework of catastrophe theory .
the standard approach allows to obtain the canonical expression for the normal form @xmath39 justifying the existence of cusp catastrophe in the problem ( fig .
[ equilibrium manifold ] ) . here , the new control variables @xmath40 and @xmath41 relate to the origin ones @xmath42 by means of relations @xmath436@xmath44 and @xmath456@xmath46 correspondingly , and @xmath476@xmath48 is the scaled phase - state variable . the direct calculations lead to the explicit form of the bifurcation set and @xmath10 ,
the system of strongly correlated electron gas is a subject for the dpt of the first order .
the only point @xmath49 where the system undergoes the second - order phase transition , e. g. the critical point , is a triple degenerate point shown by solid circle.,scaledwidth=40.0% ] as a function of magnetic induction for fix value @xmath72.25 and three different value of @xmath90 , 0.5 and 1.5 is shown for one dhva period .
( a ) for @xmath90 and @xmath501 , the system has twofold degenerate ground state : two minimums ( circles ) corresponding to the same value of @xmath51 can give rise to the simple plane - parallel domain structure .
( b , c ) the increase of @xmath10 followed by increase of @xmath52 breaks the equivalency of the steady states , results in increase of the difference between the the energies of two equilibrium states and possibility of realization of variety of domain structures . for @xmath72.25 and @xmath91.5 the only minimum implies the absolute stability of uniform diamagnetic phase .
, scaledwidth=40.0% ] @xmath53 defining the bifurcation curve on @xmath42 plane of control variables where the fold bifurcation occurs . eq .
( [ eq : bifurcation set 1 ] ) is a generalized condition for dpt occurrence .
it can be simplified in the vicinity of triple degenerate point ( a=1,x=0 ) resulting in standard cusp bifurcation set @xmath54 which is a semicubic parabola .
it follows from eq .
( [ eq : bifurcation set 1 ] ) that the usually accepted condition for dpt occurrence @xmath71 @xcite is justified only at the center of the dhva period , @xmath15 , but violated for @xmath550 and replaced by @xmath56 where @xmath52 is defined by eq .
( [ eq : bifurcation set 1 ] ) is a field - dependent . in particular , for @xmath57 0 , we obtain @xmath58 the character of bifurcation may be seen from the surface plot @xmath37 , or the surface of stationary solutions of @xmath59 .
the results of calculations are represented in fig .
[ equilibrium manifold ] which shows the equilibrium manifold of the system of strongly correlated electron gas for one period of dhva oscillations in joint space of state and control variables @xmath60 and the bifurcation set eq .
( [ eq : bifurcation set 1 ] ) in @xmath42-plane . at the center of dhva period ( @xmath15 ) and additional condition @xmath61
the system exhibits the symmetric pitchfork bifurcation and undergoes the phase transition of the second order from homogeneous to inhomogeneous state at decrease of temperature @xmath62 ( increase of @xmath1 ) with formation of cd structure the type of which is governed by the long - range dipole - dipole interaction and depends on the shape of the sample .
the critical point @xmath63 , e. g. triple degenerate point , is the only point where the dpt is of the second order .
but , at @xmath36 , a discontinues jump of order parameter implies the possibility of existence of the first order dpt when the projection of equilibrium state on @xmath42-plane crosses the bifurcation set with change of temperature @xmath62 or magnetic field @xmath10 ( see , also fig .
[ magnetization ] ) .
a simple graphical illustration of above - mentioned effects is provided by corresponding potential function , @xmath64 , drawn as a function of magnetic induction @xmath65 for @xmath7 2.25 and three different value of @xmath10 in fig .
[ potential function ] which shows the possibility of formation of cd structure due to existence of two minima around the center of dhva period ( fig .
[ potential function]a , b ) and the absence of the condition for realization of idp for greater values of @xmath10 when two minima coalesce and the uniform ground state is not degenerate ( fig . [ potential function]c ) . , and
( b ) internal magnetic fields , @xmath10 , at crossing the bifurcation set from high field side are plotted as a functions of applied field , @xmath31 , for different values of demagnetizing coefficient , @xmath66 . at given value of internal field , shown by horizontal dash line in ( b ) , the values @xmath67 ( @xmath681 - 6 ) of applied field correspond to the increasing values of @xmath66 .
both function , @xmath32 and @xmath10 , are odd functions of @xmath31 , thus , only positive interval of @xmath31 is shown .
, scaledwidth=40.0% ] the catastrophe theory allows to eliminate all the possible stable equilibria ( local minima ) for a given pair of control variables ( @xmath69 ) with the magnetic induction @xmath65 , as a phase - state variable , and establish the boundaries of cds instability .
thus , the straightforward application of the results consists in evaluation of the range of existence of idp , @xmath70 , defined as a part of dhva period occupied by cds .
the boundary between different phases is governed by the bifurcation set of the cusp catastrophe . due to the long - range character
, the dipole - dipole interaction is sensitive to the shape of a sample which results in the dependence of the internal magnetic field @xmath10 on experiment arrangement .
the simplest way to account for the shape of the sample is the use of demagnetizing coefficient @xmath66 for uniformly magnetized ellipsoid subjected to the uniform magnetic field along principal axis . in this case , the magnetization and magnetic induction are single - valued functions of applied magnetic field , @xmath31 , the ground state of the system is uniform and the shape effects can be calculated numerically .
this procedure allows to evaluate the conditions of occurrence of the diamagnetic instability .
of cause , the morphology of the domain patterns in the idp can not be account for by the use of the conception of demagnetizing coefficient , but has to be established by solving maxwell equations .
the results of calculation of the stray field , @xmath32 , as well as the internal field , @xmath10 , as a function of applied field , @xmath31 , for several values of @xmath66 when the uniform magnetization goes to the value defined by bifurcation set eq .
( [ eq : bifurcation set 1 ] ) are plotted in fig .
[ internal vs applied fields ] which illustrates the well - known effect , e. g. the increase of @xmath66 results in the growing difference between internal and applied magnetic fields due to the increase of absolute value of stray fields .
in particular , considering the samples with different demagnetizing coefficients , it follows from fig .
[ internal vs applied fields](b ) , that at other equal conditions , the phase boundary is reached at the higher values of applied field for the sample with the larger value of @xmath66 .
therefore , idp occupies the larger part of the dhva period for the sample with @xmath71 which are preferable for observation of cd instability . in connection to this
, it is worth to be mentioned that the discovery of the condon domains was made on a plate - like sample of ag @xcite and the diamagnetic instability was observed mostly by use the sample with the values of relevant demagnetizing coefficient in the range @xmath72 0.5 - 1 ( see , @xcite , @xcite ) . the effect of `` broadening '' of idp when the phase boundary is drawn in ( @xmath73 ) plane is illustrated by fig .
[ phase boundary ] which shows the first - order phase curves for several values of demagnetizing coefficient . in the calculation we neglected the overheating - undercooling effects . in one period of dhva oscillations .
at given value of @xmath66 , the inner ( outer ) region , designated by @xmath74 ( @xmath75 ) , corresponds to non - uniform ( uniform ) diamagnetic phase . for fix @xmath1 ( horizontal dash line )
the part of dhva period occupied by idp is wider for the sample with larger demagnetizing coefficient.,scaledwidth=40.0% ] to compare the results of the theory with available experimental data on observation of cds instability we calculate the characteristics of idp .
the formation of idp is characterized by two measured principal parameters , the average magnetic induction splitting , @xmath76 , defined as the difference , calculated at the center of dhva period , between the values of magnetic induction in two adjacent domains and the range of existence of cds , @xmath77 , defined as a part of dhva period occupied by non - uniform phase @xcite .
though the occurrence of idp and the existence of cds itself depends on long - range dipole - dipole interaction , the average value of magnetic induction splitting , @xmath76 , depends on the differential magnetic susceptibility and with reasonable accuracy is defined in a explicit form by @xcite @xmath78 as far as concerned the range of existence of idp @xmath77 , it is a shape - dependent quality due to dependence of internal field on dipole - dipole interaction sensitive to the sample shape , as illustrated in fig .
( [ internal vs applied fields ] ) .
inasmuch as both characteristics of idp originate from the same eq .
( [ eq : magnetization ] ) , they can be related to each other by eliminating @xmath1 .
it results in rather complicated functional dependence which is handled numerically , but in a limit @xmath79 0 has a simple form of linear function @xmath80 from mathematical point of view eq .
( [ eq : slope ] ) is fulfilled in an infinitely small vicinity of the triple degenerate point ( @xmath90 , @xmath7 1 ) , justifying the applicability of shoenberg theory of shape effect at this critical point .
accounting for the shape of the sample we calculate the family of the curves @xmath81 plotted in fig .
[ idp parameters ] .
the graphical representation of function @xmath82 which relates the characteristics of idp to the shape of the sample is convenient to use for comparing the theoretical results with experiment .
for this , we introduce also the characteristics of idp which can be measured ( or evaluated ) directly in experiment , @xmath83 and @xmath84 .
is shown for one dhva period .
the slope of the curves at @xmath85 is defined by inverse demagnetizing coefficient , 1@xmath86 , in accordance with eq .
( [ eq : slope ] ) justifying the shoenberg theory of shape effects at the vicinity of critical point ( @xmath1=1 , @xmath10=0 ) .
the symbols accompanied by reference number represent the experimental data.,scaledwidth=40.0% ] the formation of idp is accompanied by irreversible behavior of the magnetization curve which can be used for experimental investigation of idp .
the experimental detection of diamagnetic hysteresis was performed on be single crystal rod - like sample of size 8@xmath872@xmath871 mm@xmath88 @xcite . the prolate ellipsoid which approximates the sample has demagnetizing coefficient @xmath89 0.04 .
the hysteresis was observed at @xmath900.25 part of the dhva period which is 2@xmath13 in reduced units .
it is reasonable to associate the range of existence of idp with the range of existence of hysteresis .
thus , we obtain @xmath911.57 .
unfortunately , the magnetic induction splitting is not reported in experiment , but it can be calculated by use of the value @xmath1 due to eq .
( [ eq : splitting ] ) . at the conditions of experiment ,
@xmath61.3 k , @xmath922 k ( dingle temperature ) and @xmath933.6 t , the differential magnetic susceptibility calculated in the model of slightly corrugated fermi surface @xcite is @xmath942 which with the aid of eq .
( [ eq : splitting ] ) gives @xmath953.8 . to verify the consistency of this value of @xmath76 we calculate the absolute value of induction splitting @xmath9678.6 g using the reported value of dhva period @xmath97130 g and compare it with the experimentally detected global change of induction @xmath98121 g at one dhva period .
the inequality @xmath99 confirms the reliability of calculated value of @xmath1003.8 . the point corresponding to the values of @xmath911.57 and @xmath1003.8 is plotted in fig .
[ idp parameters ] and lies at the curve for @xmath1010.04 , as it is expected .
the series of experiments on observation of cds by @xmath2sr spectroscopy @xcite-@xcite have proved the occurrence of diamagnetic instability of strongly correlated electron gas in a number of normal metals , revealed the possibility of the first - order phase transition by detecting the overheating - undercooling effects and actually served as one of the stimulation factors for development of the theory . as an example , we consider the data for be @xcite which provide the value of magnetic induction splitting and range of existence of cds with high accuracy of @xmath1021 g in applied field of several tesla .
the @xmath2sr measurements in @xcite were performed on plate - like samples of size 1@xmath871@xmath878 mm@xmath88 . at the conditions of experiment , @xmath60.8 k and @xmath932.75 t , the experimental value of induction splitting @xmath103 [email protected] g which gives @xmath105 2.31 with use of the reported dhva period @xmath10678.2 g. the value of the range of existence of idp is @xmath10727 g or @xmath912.2 .
the plotted in fig .
[ idp parameters ] point , @xmath912.2 and @xmath1002.31 , belongs to the curve with @xmath720.8 which is close to the theoretical value of @xmath890.75 for oblate ellipsoid approximating the sample . the famous experiment on observation of cds instability by nmr technique @xcite on the plate - like sample of ag of size 8@[email protected] mm@xmath88 at helium temperature and applied field
@xmath939 t provides both values @xmath108@xmath109 11 which allows to calculate the reduced values @xmath110@xmath111 4.32 with use of experimental value of dhva period @xmath10615.9 g. the corresponding point plotted in fig .
[ idp parameters ] lies at the curve parameterized by @xmath101 0.78 which is close to the value @xmath72 0.84 obtained for inscribed oblate ellipsoid approximated the sample .
thus , there is a good agreement of the theoretical results with the experimental data . the absence of data at the vicinity of origin ( fig . [ idp parameters ] )
definitely calls for further experimental investigation of diamagnetic instability of strongly correlated electron systems .
one of the properties of the family of curves @xmath81 fig .
[ idp parameters ] consists in its universality in a sense that the curves being a graphical representation of the relationship between the measured properties of idp ( order parameter @xmath76 and range of existence of idp @xmath77 ) and characteristics of the shape of the sample ( demagnetizing coefficient @xmath66 ) do not depend on concrete fermi surface .
therefore , the data related to different systems exhibiting the diamagnetic instability can belong to the same curve . one can conclude that plotted in the ( @xmath112 ) plane where @xmath113 is the range of existence of idp in terms of internal magnetic field , the family of curves ( fig .
[ idp parameters ] ) gathers onto a single universal , or parent curve @xmath114 .
this is a consequence of the universality of the bifurcation theory in description of diamagnetic instability of the electron system in a joint space ( b , a , x ) of phase and control variables .
another feature of the dependencies represented in fig .
[ idp parameters ] consists in a possibility of calculation the value of @xmath1 for given experimental arrangement . as it follows from eq .
( [ eq : splitting ] ) , a straight horizontal line in fig . [ idp parameters ] crosses the family of curves in points of equal @xmath1 .
mapping the points belonging to different curves onto the parent curve @xmath115 parameterized by @xmath1 , with careful tabulation provides the values of @xmath1 which can be compared with corresponding values calculated in different models of fermi surface .
some examples of the mapping of the experimental points onto parent curve are shown in fig .
[ idp parameters ] by dash dot lines .
it is follows from the analysis of the function @xmath116 that predicted dependency on the shape of the sample can be detectable for moderate values of reduced amplitude of dhva oscillations @xmath94 1.3 - 3 . for greater values of @xmath1
the high precision experiments are needed due to the small difference between the curves .
the diamagnetic instability and stable equilibria of the system of the strongly correlated electron gas at high magnetic field and low temperature are studied by the tools of catastrophe theory . while at the center of dhva period the symmetric supercritical pitchfork bifurcation gives rise to the second order phase transition on temperature , the deviation of magnetic field from the value corresponding to the center of dhva period results in the change of the phase transition order from the second to the weakly first in the nearest vicinity of triple degenerate point and first at the rest part of dhva period both on temperature and magnetic field .
we calculate the generalized condition for dpt occurrence .
the results of our investigation show a reasonable agreement with the experiments .
we believe also that the observation of dpt and interpretation of data can be complicated by the intrinsic structure of idp which can realized through band patterns , bubbles and so on , but this is beyond the scope of the present publication .
we hope that our studies provide consistent framework in consideration of diamagnetic instability of strongly correlated electron gas , will help in making a quantitative systematic analysis of experimental data and stimulate the further experiments on investigation of the diamagnetic instability .
j. h. condon , phys .
45 , ( 1966 ) 526 .
j. h. condon and r. e. walstedt , phys .
21 , ( 1968 ) 612 .
r. b. g. kramer , v. s. egorov , v. a. gasparov , a. g. m. jansen , and w. joss , phys .
95 , ( 2005 ) 267209 .
r. b. g. kramer , v. s. egorov , a. g. m. jansen , and w. joss , phys .
95 , ( 2005 ) 187204 .
m. i. tsindlekht , n. logoboy , v. s. egorov , r. b. g. kramer , a. g. m. jansen , and w. joss , phys . of low temp .
32 , ( 2006 ) 1129 .
g. solt , c. baines , v. s. egorov , d. herlach , e. krasnoperov and u. zimmermann , phys .
76 , ( 1996 ) 2575 .
g. solt , c. baines , v. s. egorov , d. herlach , and u. zimmermann , phys .
59 , ( 1999 ) 6834 .
g. solt and v. s. egorov , physica b 318 , ( 2002 ) 231 .
n. logoboy , a. gordon , i. d. vagner , and w. joss , solid state comm .
134 , ( 2005 ) 497 .
n. logoboy and w. joss , solid state comm .
139 , ( 2006 ) 240 .
n. logoboy and w. joss , solid state comm .
139 , ( 2006 ) 191 .
i. m. lifshitz and a. m. kosevich , sov .
jetp 2 , ( 1956 ) 636 .
d. shoenberg , _ magnetic oscillations in metals _
( cambridge university press , cambridge , england , 1984 ) .
r. d. plummer and w. l. gordon , phys
13 , ( 1964 ) 432 .
j. d. livingstone and w. desorbo , in superconductivity , edited by r. d. parks ( marcel dekker , new york 1969 ) .
a. hubert and r. sch@xmath117fer , _ magnetic domains - the analysis of magnetic microstructures _ ( springer - verlag , heidelberg , 2002 ) .
b. binz , h. b. braun , t. m. rice , and m. sigrist , phys .
96 , ( 2006 ) 196406 .
m. blanter , m. i. kaganov , and d. v. posvyyanskii , phys .
38 , ( 1995 ) 203 . | in a framework of catastrophe theory we investigate the equilibrium set for the system of strongly correlated electron gas at the conditions of dhva effect and show that the discontinuities accompanied the diamagnetic phase transition ( dpt ) is handled by riemann - hugoniot catastrophe .
we show that applicability of the standard condition for observation of dpt @xmath01 where @xmath1 is the differential magnetic susceptibility is valid only in the nearest vicinity of triple degenerate point corresponding to the center of dhva period , but for arbitrary value of magnetic field in every period of dhva oscillations this condition is modified in accordance with the bifurcation set of cusp catastrophe . while at the center of dhva period the symmetric supercritical pitchfork bifurcation gives rise to the second order phase transition on temperature , the deviation of magnetic field from the value corresponding to the center of dhva period results in the change of the phase transition order from the second to the first one both on temperature and magnetic field . in the framework of developed theory
we obtain good agreement with available experimental data . |
the kolmogorov - johnson - mehl - avrami ( kjma ) model finds application in a vast ambit of scientific fields which ranges from thin film growth to materials science to biology and pharmacology , let alone the applied probability theory . in the majority of these studies
the authors made use of a simplified version of the kjma formula : the stretched exponential @xmath0 , where @xmath1 is the fraction of the transformed phase , @xmath2 and @xmath3 ( the latter known as avrami s exponent ) being constants .
the model , in principle , is simple because it rests on a poissonian stochastic process of points in space , to which a growth law is attached .
in fact , owing to the poissonian process , the nucleation takes place everywhere in the space i.e. also in the already transformed phase . this partially fictitious nucleation rate ( @xmath4 ) , for we are dealing with a poissonian process , is linked to the actual ( real ) nucleation rate ( @xmath5 ) according to : @xmath6 $ ] , where @xmath7 is the transformed fraction .
the growth law transforms each point in a nucleus of radius @xmath8 , @xmath9 stands for time .
the pair , `` points generation '' and `` growth law '' , is a key quantity of the theory .
it happens that the kjma model fails for time - dependent points generation rate ( i.e. nucleation rate ) associated with diffusional - type growth laws @xcite,@xcite .
in particular , let us define two classes of growth laws : i ) @xmath10 and ii ) @xmath11 .
the kjma model is suitable for describing the first class of growths and for this reason we named it kjma - compliant as opposite to the second to which we attach the adjective kjma - non - compliant .
the reason for that is due to the particular stochastic process taken into account . as a matter of fact
, the poissonian process requires that points can be generated everywhere throughout the space independently of whether the space is , because of growth , already transformed or not .
points generated in the already transformed space are named phantoms after avrami .
it goes without saying that , in the case of kjma - compliant growths , phantoms do not contribute to the true transformed fraction , i.e. they are just virtual points whose only role is to simplify the mathematics @xcite . on the other hand , in the kjma - non - compliant growths phantoms
may contribute to the phase transition through the non - physical `` overgrowth '' events @xcite .
incidentally , it is worth noticing that the kjma - non - compliant growths and kjma - compliant growths are indistinguishable if associated with simultaneous nucleation @xcite . according to what has been said ,
one can summarize saying that the concept of phantom implies the existence of the two classes of growths .
one has to bring in mind that the above stretched exponential expression ( @xmath0 ) is the exact solution of the kinetics only in the case @xmath12 and @xmath13 , @xmath14 being the nucleation rate and @xmath15 dirac s delta function , provided the growth is according to a power law . in general , the term @xmath16 is a simple way to approximate the convolution product between the nucleation rate ( phantom included ) and the nucleus volume .
this convolution is the `` extended '' transformed fraction and takes into account the contribution of phantoms , @xmath17 . in view of the large use of the kjma theory for dealing with experimental data
, it should be desirable to make use of an extended transformed fraction deprived of phantom contribution , @xmath18 .
a pictorial view of the geometrical meaning of @xmath17 , @xmath18 and @xmath1 is reported in fig.1 . ) and actual ( @xmath18 ) extended fractions are also reported .
the ( extended ) volume of the i - th nucleus is denoted as @xmath19 and the transformed fraction as @xmath1 . ]
the aim of this contribution is twofold : i ) to model the phase transformation kinetics in terms of actual quantities such as the nucleation rate ; ii ) to provide an expression for the transformed fraction as a function of @xmath18 .
in this section we discuss the stochastic theory that has to be employed in order to get rid of phantoms in modeling the kinetics of phase transformations ruled by nucleation and growth . to this end , let us define the phantom included nucleation rate @xmath4 , and the `` actual '' nucleation rate , @xmath5 , namely the rate of birth of the `` real '' nuclei ; @xmath5 is the quantity experimentally accessible .
it goes without saying that a mathematical formulation of the phase transition kinetics which employs the `` actual '' nucleation rate holds true for both kjma - compliant and kjma - non - compliant growth laws .
this automatically overcomes the limit of the kjma approach . however , actual nuclei imply a severe complication of the stochastic - nature of the process under study , for we shift from a poissonian to a non - poissonian process : in fact actual nuclei are spatially correlated
. let us address this point in more detail by denoting with @xmath20 the radius of a nucleus , at running time @xmath9 , which starts growing at time @xmath21 . to be an actual nucleus it has to lie at a distance @xmath22 from any other `` older '' nucleus with @xmath23 .
in other words , in the spirit of the statistical mechanics of hard spheres , this condition is formalized ( at the lowest order ) through the pair distribution function for the `` pair of nuclei ( @xmath24 ) '' at relative distance @xmath25 @xmath26 where @xmath27 is the heavyside function . throughout the paper
we employ the notation by van kampen @xcite according to which @xmath28dots distribution and correlation functions are denoted as @xmath29 and @xmath30 , respectively . in previous papers ,
we have presented a theory for describing phase transitions in the case of spatially correlated nuclei and for time dependent nucleation rate .
the untransformed fraction can be expressed in terms of either the distribution functions , ( @xmath29-functions ) , or the correlation functions , ( @xmath30-functions ) , where the @xmath29 s and @xmath30 s are linked by cluster expansion @xcite .
given a generic point of space we have computed the probability that this point is not covered ( transformed ) by any nucleus up to time @xmath9 .
this probability is the fraction of untransformed phase i.e. @xmath31 . by denoting with @xmath32 the volume of a nucleus , which starts growing at time @xmath21 , at running time @xmath9 , and in the case of symmetric @xmath29 and @xmath30 functions ,
the uncovered fraction is given by @xmath33 or , @xmath34.\end{aligned}\ ] ] it is worth pointing out that the nucleation rates entering these equations are in fact subjected to the condition imposed by the correlation among nuclei .
this quantity may or may not imply phantoms depending on the specific form of the @xmath29 functions .
for this reason we introduce the new symbol @xmath35 .
in particular , for the hard core correlation ( eqn.[f1 ] ) @xmath35 coincides with the actual nucleation rate , @xmath36 , which leads to the solution of the phase transition kinetics in terms of the actual nucleation rate . since eqn.[uncov1],[uncov2 ] are the exact solutions of the stochastic process linked to the phase transition
, they also coincide with the kjma formula provided the above mentioned preconditions are met i.e. random nucleation and kjma - compliant growth . as far as the hard - core correlation ( eqn.[f1 ] ) and the kinetics eqns.[uncov1],[uncov2 ] are concerned , we note that the number of nuclei of size @xmath37 is @xmath38 . as a consequence and in the framework of the statistical mechanics of hard sphere fluid , we are dealing with an extremely dilute solution of pairs of components @xmath39 .
accordingly , the number density of @xmath40 spheres being of the order of @xmath41 , higher order terms in the cluster expansion of the @xmath42 function can be neglected , thus @xmath43 in eqn.[uncov1 ] . in this section
we show that eqn.[uncov1 ] is compatible with the kjma kinetics only in the case of kjma - compliant function . on the other hand , such a comparison will give a deeper insight into the reasons why the kjma kinetics does not work in the case of `` kjma - non - compliant '' functions .
in the following , we discuss the linear growth law for 2d case .
also , to simplify the complexity of the computation the `` actual '' nucleation rate , @xmath44 , is taken as constant ; as a consequence the phantom included nucleation rate reads @xmath45 and the kjma kinetics becomes @xmath46,\ ] ] where @xmath47 and @xmath48 is a constant .
we consider the series expansion of @xmath49 around @xmath50 .
one gets , @xmath51 where @xmath52 . moreover ,
since @xmath53 , being @xmath54 and @xmath55 , from eqn.[serie1 ] it is found that only the terms @xmath56 are different from zero , i.e. @xmath57 . in particular ,
the first coefficients are : @xmath58 .
next we derive the series expansion of the untransformed fraction @xmath59 , by exploiting the condition @xmath60 . even in this case
the @xmath61 coefficients are different from zero for @xmath62 ( with integer @xmath63 ) .
the first four coefficients are @xmath64 by using the expression of @xmath65 s the series of the untransformed fraction up to @xmath66 is given as @xmath67 since @xmath68 the series eqn.[serie2 ] can be rewritten as @xmath69 the two last coefficients being @xmath70 and @xmath71 .
we emphasize that the coefficients of this series only depend upon growth law , for constant @xmath44 ( see also the last section ) .
the next step is to show that the untransformed fraction , given by the series eqn.[uncov1 ] is equal to eqn.[serie2 ] or eqn.[serie3 ] .
we have carried out the first two terms , exactly , while , owing to the tremendous computational complexity , for the third term an approximation has been employed .
it is worth reminding that the distribution functions are @xmath72 and @xmath73 , where @xmath74 is the relative distance .
in fact , since @xmath75 the system of dots is dilute and @xmath76 i.e. the superposition principle holds true @xcite . since the system is homogeneous
the @xmath29 functions depend on @xmath77 .
eqn.[uncov1 ] becomes @xmath78 where the integration domain @xmath79 is the circle of radius @xmath80 .
the @xmath81 containing term is the extended surface fraction @xmath82 and coincides with the second term of the expansion eqn.[serie3 ] .
let us focus our attention on the integrals in the spatial domain - for the sake of clarity shown in fig.2a together with the circle of correlation @xmath83- of the @xmath42 containing term . by employing relative
coordinates the integrals read function .
the cases of kjma - compliant and kjma - non - compliant growths are reported in panels a and b , respectively . in the drawing @xmath84 , @xmath85 ( @xmath86 ) and @xmath87 . in the case of kjma - compliant growth @xmath88 and the hard core circle
is within the integration domain , @xmath89 ( a ) . in the case of kjma - non - compliant growths @xmath90 and the hard core disk overcomes the integration domain of the `` older nucleus '' ( b ) . ]
@xmath91 where @xmath92 is the area `` spanned '' by the second nucleus when the first one is located at @xmath93 ( @xmath94 ) .
it is at this point of the computation that the growth law comes into play ; indeed in the case of kjma - compliant growth laws ( here linear growth ) the correlation circle @xmath83 is entirely within the integration domain @xmath95 ( fig.2a ) .
consequently , @xmath96\ ] ] is independent of @xmath93 . on the other hand , in the case of kjma - non - compliant growth laws
the correlation circle overcomes the integration domain of the second nucleus and the relationship above does not hold true , anymore ( see fig.2b ) .
in general , the kjma - compliant functions satisfy the condition @xmath97 i.e. for a power growth law @xmath98 , which is verified only for @xmath99 .
in fact , by setting @xmath100 and @xmath101 , the inequality above reads @xmath102 , which is satisfied for @xmath99 .
on the other hand , for @xmath103 ( with integer @xmath104 ) the inequality is @xmath105^{k}\leq \eta$ ] namely , @xmath106 which is never satisfied ( @xmath107 ) . for the kjma - compliant growth
the contribution of the @xmath42 containing term becomes @xmath108\\ \\
= \pi^{2}i_{a}^{2}\int_{0}^{t}dt_{1}\int_{0}^{t_{1}}dt_{2}r^{2}(t , t_{1 } ) \big [ r^{2}(t , t_{2})-r^{2}(t_{1},t_{2})\big]\\ \\ = \frac{x_{e}^{2}}{2}-\pi^{2}i_{a}^{2}\int_{0}^{t}dt_{1}\int_{0}^{t_{1}}dt_{2}r^{2}(t , t_{1})r^{2}(t_{1},t_{2 } ) .
\end{array}\ ] ] it is possible to show that for linear growth the last term of eqn.[secondterm ] is equal to @xmath109 , consequently we get @xmath110 , which coincides with the term of the same order in the kjma series eqn.[serie3 ] . let s now briefly consider the contribution of the @xmath111 containing term in the general expression eqn.[uncov1 ] .
fig.3 shows the integration domains @xmath112 and the correlation circles @xmath113 for the three nuclei born at time @xmath114 ( @xmath115 ) .
the configuration integral over @xmath111 in eqn.[coeffic1 ] becomes function . in the drawing @xmath116 and @xmath117 . in the case of kjma - compliant growth
all the circles are entirely within the @xmath118 circle . ]
@xmath119 where @xmath120 is assumed . in this equation , @xmath121 is the area spanned by the third nucleus when the first and the second are located at @xmath93 and @xmath122 , respectively .
because of the possible overlap between the correlation circles @xmath123 and @xmath124 , and for @xmath123 is encompassed within @xmath118 this area is a function of @xmath125 as @xmath126h(r_{12}-r(t_{2},t_{3})-r(t_{1},t_{3 } ) ) , \end{array}\ ] ] where @xmath127 , with @xmath128 being the overlap area of two circles of radius @xmath129 and @xmath130 at relative distance @xmath131 @xmath132^{2}}+\\ \\ + \rho_{1}^{2}\arccos \frac{\rho_{1}^{2}+x^{2}-\rho_{2}^{2}}{2x\rho_{1}}+ \rho_{2}^{2}\arccos\frac{\rho_{2}^{2}+x^{2}-\rho_{1}^{2}}{2x\rho_{2}}. \end{array}\ ] ] it turns out that the computation of the third order term of the series , eqn.[f3 ] , is a formidable task indeed .
we do not attempt to perform the exact estimate of this term which , however , must coincide with the same order term of the kjma series . on the other hand ,
an approximate evaluation of this term by using an oversimplified form of the @xmath133 area , is possible by formally rewriting this area as @xmath134 $ ] , where @xmath135 is given by @xmath136 $ ] . in the case of complete - overlap ( @xmath137 )
we get , @xmath138[r^{2}(t , t_{3})-r^{2}(t_{1},t_{3 } ) ] , \end{array}\ ] ] that for the linear growth ( @xmath47 ) gives @xmath139 to be compared with the exact value @xmath140 which brings an uncertainty of @xmath141 .
let us now consider the parabolic growth @xmath142 . in this case
the series expansion of the function @xmath143 , given by eqn.[kjma2 ] , can be performed by employing the same computation pathway discussed above , where now @xmath144 and @xmath145 . in this case
@xmath146 implying @xmath147 and @xmath148 the two last coefficients being @xmath149 and @xmath150 .
it is worth pointing out that in such an evaluation the transformed fraction , @xmath1 , is comprehensive of the contribution of phantoms .
in fact , we recall that eqn.[kjma2 ] is the kjma solution with the phantom included nucleation rate , @xmath151 . the @xmath81 containing term of eqn.[uncov1 ] gives the extended volume fraction @xmath152 .
as far as the third term is concerned ( @xmath42 contribution ) , it is possible to show that also for @xmath153 the integral eqn.[secondterm ] coincides with the third term of eqn.[parabola1 ] , @xmath154 .
however , it is important to stress that eqn.[secondterm ] does not coincide , in this case , with the integral over the @xmath42 function of the exact solution eqn.[coeffic1 ] , since in the latter equation enter the actual nuclei , only . from the mathematical point of view , in the case of parabolic growth , the correlation circle is not contained within the integration domain as depicted in fig.2b .
in other words , for kjma - non - compliant growth laws the area @xmath155 is a function of @xmath156 and the term of order @xmath157 in eqn.[coeffic1 ] does not coincide with @xmath158 of eqn.[parabola1 ] ( parabolic growth ) . in particular , under these circumstances
we get @xmath159h(r(t , t_{2})-r_{1}-r(t_{1},t_{2 } ) ) , \end{array}\ ] ] where @xmath160 with @xmath161 the overlap area of two circles of radius @xmath95 and @xmath83 at distance @xmath156 ( eqn.[areoverlap ] ) .
the ultimate aim of this section is to propose a simple formula for describing the kinetics on the basis of the `` actual '' extended transformed fraction , @xmath18 . on the ground of eqn.[uncov2 ]
the transformed fraction can be rewritten in the general form @xmath162,\ ] ] where @xmath163 embodies the contributions of correlations among nuclei @xcite .
it is worth noticing that for kjma - complaint growths ( random nucleation ) eqn.[uncov2 ] actually reproduces the kjma formula .
in fact by identifying @xmath35 with the phantom included nucleation rate ( @xmath164 ) one gets @xmath165 leading to the formula @xmath166 . on the other hand ,
working with the actual nucleation rate , in eqn.[uncov2 ] @xmath167 , @xmath168 and the series has infinite terms .
@xmath169 can therefore be expanded as a power series of the extended actual volume fraction , @xmath18 . moreover , by exploiting the homogeneity properties of the @xmath29 functions ( see below ) , it is possible to attach a physical meaning to the power series coefficients , in terms of nucleation rate and growth law .
also , for constant @xmath44 the coefficients of this series only depend upon growth law .
to the aim of achieving a suitable compromise between handiness and pliability , we retain the liner approximation of @xmath170 obtaining the following kinetics @xmath171,\ ] ] with @xmath2 and @xmath172 constants . for the sake of completeness we point out that , according to its physical meaning , the parameter @xmath2 should be unitary .
nevertheless , the substitution of the infinite expansion with only two terms authorizes the introduction of the new parameter @xmath2 . in any case
, the @xmath2 values are found to be nearly one ( see fig.7 below ) . in order to study the transition kinetics in terms of actual nuclei and to test eqn.[sim2 ] , we worked out 2d computer simulations for several growth laws at constant nucleation rate , @xmath44 .
as typical for this kind of study @xcite , @xcite , the simulation is performed on a lattice ( square in our case ) where , in order to mimic the continuum case , the lattice space is much lower than the mean size of nuclei .
in particular , the transformation takes place on a square lattice whose dimension is @xmath173 with a nucleation rate of @xmath174 .
it is worth reminding that , since the nucleation is poissonian , it occurs on the entire lattice independently of whether the space is already transformed or not .
the computer simulation can be run taking into account the presence of phantoms or not . in the former case
the outputs have been labeled as `` w '' while the latter as `` wo '' .
as far as the growth laws are concerned , we limited ourself to the power laws , @xmath175 for @xmath176 and @xmath177 .
the results of the simulations are displayed in figs.4a - c for the kjma - non - compliant growths .
in particular , the fractional surface coverage,@xmath1 , as a function of the actual extended fraction @xmath18 , with and without the contribution of phantoms , are reported ( curves labeled with `` w '' and `` wo '' , respectively ) . , is shown as a function of the extended fraction @xmath18 , for the power law ( @xmath178 ) where @xmath153 , @xmath179 and @xmath180 in figures a , b and c , respectively .
the kinetics with ( w ) and without ( wo ) the inclusion of phantoms are displayed . ]
the contribution of phantom overgrowth to the transformation kinetics is highlighted in fig.5 and shows that this effect brings an uncertainty on @xmath1 which ranges from @xmath181 to @xmath182 on going from @xmath153 to @xmath183 . in the case of parabolic growth
this figure is lower than @xmath181 . ,
@xmath184 and @xmath185 for @xmath183 , @xmath186 and @xmath153 . ]
these results are in qualitative agreement with previous studies on phantom overgrowth , although performed for a different nucleation laws .
as discussed in more details below , the results displayed in figs.4,5 are universal , i.e. they only depend on power exponent , @xmath3 , and nucleation law ( in the present case @xmath187 ) .
accordingly , the lower @xmath3 the more important is phantom overgrowth .
in fact , let us consider a phantom , which starts growing at time @xmath188 , located at @xmath189 from the center of an actual nucleus which starts growing at @xmath50 ( fig.6a ) . for kjma - non - compliant growth , @xmath190 ( with integer @xmath104 ) , the phantom overtakes the actual nucleus at time @xmath191 , that is the solution of the equation @xmath192 , namely denotes the location of the phantom which start growing at time @xmath188 when the size of the actual nucleus is @xmath193 .
the phantom overtakes the actual nucleus at time @xmath191 when the size of the actual nucleus is @xmath194 . ]
@xmath195 where @xmath196 and @xmath197 .
the graphical solution of eqn.[ovegrowth ] is depicted in fig.6b and indicates that @xmath191 ( and therefore @xmath198 ) decreases with @xmath63 .
this is in agreement with the results of fig.5 which shows that the overgrowth phenomenon is more important at greater @xmath63 .
as far as the guess function eqn.[sim2 ] is concerned , it matches the simulation curves with a very high degree of correlation .
for instance the output of the fit to the @xmath199 curve , gives @xmath200 , @xmath201 and a squared correlation coefficients practically @xmath202 .
for the sake of completeness the @xmath2 and @xmath172 fitting parameters are shown in fig.7 , where @xmath2 is found to be nearly one .
this is in agreement with the theoretical value predicted by eqn.[uncov2 ] . as a function of growth exponent @xmath3 , for kjma - non - compliant growths . ]
the behavior of the transformed fraction for kjma - compliant growths are reported in fig.8 for @xmath203 , @xmath204 and @xmath199 . .
in the graph the kinetics for @xmath153 is compared with the kjma - compliant growth with @xmath203 , @xmath199 and @xmath204 ( from the top , respectively ) . in the inset
the kinetics for @xmath203 and @xmath153 are displayed together with the truncated kjma series expansions eqn.[serie3 ] and eqn.[parabola1 ] , respectively . ]
these kinetics are very close to each other and differs , markedly , from that at @xmath153 also reported in the same figure . in the inset ,
the kinetics for @xmath153 and @xmath203 are compared with the kjma series expansion eqn.[parabola1 ] and eqn.[serie3 ] , respectively .
the fact that the curves for @xmath203 , @xmath205 and @xmath177 collapse on the same curve , can be rationalized computing the coefficients of the series expansion of @xmath206 for integer @xmath3 . in particular , by employing the method discussed in the previous section the two last coefficients of the series ( e.g. eqn.[serie3 ] ) are @xmath207 , @xmath208 and @xmath209 , @xmath210 for @xmath199 and @xmath211 , respectively .
we also performed computer simulations of phase transitions for non - constant actual nucleation rate .
the output of this computation is displayed in fig.9 where the behavior of the nucleation rate is shown in the inset as function of @xmath18 .
in particular , the actual nucleation rate is given by the function @xmath212 . to the @xmath213 kinetics
has been shown as dashed line .
the correlation coefficient of the fit is in fact @xmath202 for the parameters @xmath214 and @xmath215 . the actual nucleation rate , as a function of @xmath18
, is also reported on the right scale in @xmath216 units . ] also in this case the function eqn.[sim2 ] has been found to match the kinetics with high degree of correlation where , again , the independent variable is the actual extended surface fraction .
let us address in more detail the question of the dependence of volume fraction on extended volume fraction . to this end , we discuss the second order term of the exact solution eqn.[uncov1 ] , namely @xmath217 where eqn.[area0 ] has been employed .
we point out that in eqn.[changevar1 ] @xmath218 is a second order homogeneous function of @xmath156 , @xmath219 and @xmath220 variables .
accordingly , for the growth law @xmath221 , using the re - scaled variables , @xmath222 , @xmath223 and @xmath224 the integral becomes @xmath225 eqn.[changevar2 ] takes the form @xmath226 where + @xmath227)=\frac{\int_{0}^{1}d\tau'i_{a}(\tau')\int_{0}^{\tau'}i_{a}(\tau '' ) d\tau '' \int_{\delta(\tau')}a(r_{1}',\tau',\tau'')d\mathbf{r}'_{1 } } { ( \int_{0}^{1}i_{a}(\tau')(1-\tau')^{2n}d\tau')^{2}}$ ] depends on @xmath3 and actual nucleation rate .
it is apparent that in the case discussed in the previous section @xmath228 , @xmath229 , as well as higher order coefficients , is a function of @xmath3 , only . in this case
the transformed fraction is expected of the form @xmath230 . on the other hand , in the case of
a constant `` phantom - included '' nucleation rate , @xmath231 , and eqn.[uncov1 ] becomes an integral equation for the @xmath232 unknown .
with reference to the second order term , in this case eqn.[changevar2 ] takes the general form @xmath233 which now implies the series @xmath234 ( note that this is a series expansion in terms of phantom - included extended surface ) . in the specific case of kjma - compliant growths
, however , these series reduces to the exponential series with constant coefficients @xmath235 .
it is instructive to estimate the first two coefficients in the case of linear growth . for constant phantom
included nucleation rate the untransformed fraction satisfies the integral equation , @xmath236 the first order term of this equation , namely of the order of @xmath237 , gives @xmath238 . by substituting @xmath239 in eqn.[integraleq1 ] ,
we get @xmath240 using dimensionless variables @xmath241 , @xmath223 and @xmath242 eqn.[integraleq2 ] eventually becomes @xmath243 where @xmath244 is given through eqn.[aria1 ] .
notably , the last term in eqn.[integraleq3 ] has been already estimated in eqn.[secondterm ] and is equal to @xmath245 .
the coefficient of @xmath246 is eventually computed as @xmath247 , that is the expected result .
it is worth noting that the present approach can also be applied to different convex shapes other than circles and spheres , provided the orientation of nuclei is the same ( with a possible exception for triangle ) .
this aspect has been discussed in details in refs.@xcite,@xcite .
we conclude this section by quoting the recent results of ref.@xcite . in this noteworthy contribution
the author faced the problem of describing the kinetics in terms of the actual nucleation rate .
an ingenious application of the so called `` differential critical region '' approach makes it possible to find the @xmath232 kinetics by solving an appropriate integral equation @xcite . on the other hand
, the different method employed in the present work , based on the use of correlation function , pertains to the same class of stochastic approaches on which `` kolmogorov s method '' is rooted .
it could be enlightening , for the present topic , to demonstrate that the two approaches are in fact equivalent .
we have shown that , employing the correlation function approach , the constraints on growth laws underlying the kjma theory can be eliminated . in other words , the present modeling is not constrained to any form of the growth law . the actual extended volume fraction is shown to be the natural variable of the kinetics , which implies universal curves . besides , we proposed a formula to fit experimental data by using the measurable actual extended coverage .
the displacement of the kinetics from the exponential law , i.e. the @xmath172 parameter in eqn.[sim2 ] , may give insights into the microscopic growth law of nuclei .
kolmogorov , _ bull .
urss ( cl .
_ , * 3 , 355 ( 1937 ) ; selected works of a.n .
kolmogorov , edited by a.n .
shiryayev ( kluwer , dordrecht , 1992 ) , engish translation vol.2 , pag.188 . * | the theory of kolmogorov - johnson - mehl - avrami ( kjma ) for phase transition kinetics is subjected to severe limitations concerning the functional form of the growth law .
this paper is devoted to side step this drawback through the use of correlation function approach .
moreover , we put forward an easy - to - handle formula , written in terms of the experimentally accessible actual extended volume fraction , which is found to match several types of growths .
computer simulations have been done for corroborating the theoretical approach . |
type ia supernovae ( sne ia ) are luminous events that synthesize an appreciable fraction of the iron - peak elements in the universe ( iwamoto et al . 1999 ) , and their high and relatively homogeneous peak luminosities have allowed them to be utilized to measure large distances . sn ia distance estimates have revealed that the expansion of the universe is accelerating ( riess et al .
1998 ; perlmutter et al . 1999 ) .
there is considerable interest in extending the current utilization of sne ia , both by better understanding the sn ia event and by widening the rest - frame wavelength range which can be utilized .
the ultraviolet wavelength ( uv ) range , including the range covered by the @xmath5-band , bears tremendous potential towards the better understanding of sne ia , but it also poses significant challenges . , nuv to be emission between the 2500 4000@xmath3 ( @xmath6 and @xmath7 filters ) and muv to be emission between 1500 2500@xmath3 ( @xmath8 and @xmath9 filters ) . ]
theoretically , the uv wavelength range is useful to better understand the explosive nucleosynthesis of the sn event .
iron - peak elements strongly affect this wavelength range through the related effects of line - blanketing and line - blocking ( jeffery et al .
sauer et al .
2008 , hachinger et al .
in particular , observations made at early epochs are a sensitive probe of the composition of the outer layers of the sn ejecta .
this is important for the determination of the progression of the burning front through the outer layers of the ejecta during a suggested delayed detonation phase .
the flux in the uv part of the spectrum of a sn ia is shaped mostly by metal lines , in particular feii and iii , tiii , crii , coii and niii , with contributions from siii and mgii .
these ions have a large number of strong as well as weak lines in the uv .
uv photons are actively absorbed in these lines , and because of the high velocity of the sn ejecta the forest of lines becomes a blanket " : different lines in different parts of the ejecta overlap because of velocity shear , leading to the effective blocking of the flux .
once photons are absorbed , they are most likely to escape only if they are re - emitted in red transitions , as at red ( optical ) wavelengths the total line opacity is much less .
since metal line opacity is quite high , the uv spectra are therefore good tracers of metal abundance in the outer layers of the ejecta , and different temperatures or abundance ratios can lead to different uv colours ( e.g. walker et al 2012 ) . on the other hand , in the outermost layers , uv opacity may not be so high because the density is low . in this case
the strongest metal lines are the only active ones .
these are typically in the optical .
it is therefore possible for optical photons to be absorbed and re - emitted in the uv in a process called reverse fluorescence . because of the low uv opacity in these outermost layers , reverse fluoresced photons are essential in determining the emerging uv spectrum ( e.g. mazzali 2000 ) .
hence , observations made in the uv combined with knowledge of the velocity of the expansion of the ejecta from spectra , permit detailed probing of the radiation transport on the sn ejecta .
cosmologically , the wavelength range covered by the @xmath5-band is potentially useful for distance studies .
extending a ground - based study of high redshift sne ia beyond [email protected] requires either observations made in observer - frame nir ( rest - frame optical ) , or in the observer - frame optical ( rest - frame uv ) . to explore the potential of the @xmath5-band , ground - based sn surveys have included @xmath5-band photometry in their recent campaigns of low - redshift sne .
these campaigns have reported that there is considerable scatter in the photometry , at a level that exceeds what is seen in the optical filters .
this scatter persists after all efforts to account for observational issues .
jha , riess & kirshner ( 2007 ) presented @xmath11 light curves of 44 sne ia , but found that their multi - light curve fitting algorithm had lower overall scatter without including the @xmath5-band , than it did with the @xmath5-band . one critical element for sn ia cosmology is to determine whether sn ia emission evolves with redshift .
thought to be the thermonuclear explosion of a carbon - oxygen white dwarf ( hereafter co wd ) that accretes material from a companion , evolution of emission might be caused by metallicity , differences between low and high redshift progenitor co wds or by different companions to the co wd between low and high redshift binary systems .
since modeling suggests that the uv wavelength range is a sensitive probe of the nucleosynthetic products of the sn explosion ( i.e. the composition of the ejecta , sauer et al .
2008 ; lentz et al .
2000 ) , a number of recent studies have compared rest - frame uv emission from multiple sne ia , trying to quantify the variations between events .
the rest - frame uv emission for low-@xmath12 samples have been obtained from sn ia observations made with iue ( capellaro et al .
1995 , panagia et al .
2003 ) , from observations made with hst ( jeffery et al .
1992 , see foley , filippenko & jha 2008 for a listing of all iue / hst uv spectra through 2007 ) , from high-@xmath12 samples with keck ( ellis et al .
2008 , foley et al .
2012c ) , or from hst observations of recent sne ia detected at very early epochs by the palomar transient factory ( cooke et al .
2011 , maguire et al .
2012 , hachinger et al . 2013 )
or other recent sne ia ( 2009ig : foley et al .
2012a , 2011iv : foley et al .
2012b , 2011by : foley & kirshner 2013 ) .
these studies have reported some differences between the uv emission from high-@xmath12 versus low-@xmath12 samples , and that the variations seen in the uv wavelength range within a sample exceeds the variations seen in the optical wavelength range . combined with the difficulties experienced with the ground - based @xmath5-band sn ia light curve surveys ,
the uv wavelength range has been considered a problematic wavelength range for cosmological studies .
the uvot instrument ( roming et al .
2000 , 2005 ) on the _ swift _ mission ( gehrels et al .
2004 ) possesses the ability to quickly schedule observations of a new supernova , increasing the chances of observing the important early epochs , prior to peak light .
the 30 cm uvot instrument employs 3 uv filters ( @xmath13 ) and three optical filters ( @xmath14 ) , and sne are typically observed with all six filters at all epochs .
the transmission curves for each filter are shown in figure 1 , folded through three uv / optical spectra of sne ia .
brown et al . ( 2009 ) presented photometry from the first 2.5 years of uvot observations , 2133 total observations of 25 sne , of which 17 are sne ia .
whereas hst cameras have the advantage of superior s / n photometry and spectroscopy for the sne ia that are observed , the advantage for uvot observations are superior sampling of sne ia , both in number of events and in the number of observations for a single event .
the photometry presented in brown et al .
( 2009 ) was studied for light curve properties in milne et al .
2010 ( hereafter m10 ) and for absolute magnitudes in brown et al .
2010 ( hereafter b10 ) .
m10 identified uv - optical color differences between the group of normal sne ia and the groups of subluminous and sn 2002cx - like sne ia .
the definition of normal used in m10 , which is the same definition that will be used in this work , is a sn ia with a b - band peak decline rate less than 1.6 ( e.g. @xmath15 @xmath16 1.6 ) , and not exhibiting optical spectral features similar to sn 2002cx .
this group comprises the majority of sne ia , and is generally the group utilized for cosmological distance studies . within this group of normal sne ia ,
the findings of m10 and b10 have not wholly matched the trends derived from ground - based optical and @xmath5-band surveys , nor the intepretations of the hst and iue uv spectra , as the initial sample of uvot uv light curves appeared more homogeneous than expected .
if the uv light curves are intrinsically homogeneous , or , barring that , can be organized into distinct groups and exhibit low intrinsic scatter within those groups , the uv wavelength range would prove useful for cosmological applications . in this work , we add two more years of sn ia uvot observations , to the m10/b10 collection , showing how the larger sample permits a better understanding of the range of uv light curves possible within the group of normal sne ia ( section 2 ) .
we present a more quantitative investigation into the color evolution of sub - classes we recognize within normal " sne ia .
we will not present color curves for subluminous sne ia , as those will be presented in a separate work ( milne et al .
, in preparation ) . in section 3
, we compare uvot photometry with spectrophotometry extracted from hst spectra of sne ia . in section 4
, we search for a correlation between blue uv - optical colors and optical parameters , such as the spectral signature of unburned carbon . in section 5 ,
we treat extinction and how extinction estimation needs to adapt to new findings .
new uvot photometry is presented in the appendix . in the first collective studies of uvot sn ia photometry
, m10 and b10 found that the color evolution of the uvot-@xmath6 and uv filters relative to the optical ( @xmath17 ) band was both dramatic and relatively homogeneous within the sub - class of normal sne ia .
the colors become rapidly bluer until @xmath106 days pre - peak in the @xmath18 filter , then abruptly become redder until @xmath1020 days . filter .
] there was a lone significant outlier to that trend , sn 2008q , which followed the same color - curve shape but was bluer at all early epochs than the other normal sne ia , seemingly offset from the other normal events .
the absolute magnitudes of the light curve peaks of the uvot-@xmath6 light curves were able to be fit with a linear luminosity - width relation ( lwr ) .
the scatter about that @xmath6 band lwr was similar to what was seen in the @xmath19 and @xmath17 filters .
the lwr for the @xmath20 filter had somewhat larger scatter and the @xmath21 and @xmath8 filters had large scatter .
and @xmath21 refer to red - tail corrected " @xmath7 and @xmath9 filters .
see b10 for details of the method used to try to account for contamination of those filters from longer wavelength emission .
] however , the peak pseudocolors were not correlated with peak - width for the nuv - optical colors , a finding different than what has been seen in the @xmath22 colors ( the phillips peak relation ; phillips et al .
1999 ) , or in ground - based @xmath23 colors ( folatelli et al . 2010 ) .
collectively , with the exception of sn 2008q , the uvot data - set of normal sne ia suggested that there is a high level of homogeneity in the @xmath5-band wavelength range , but significant variations at shorter wavelengths .
these findings are promising for the cosmological utilization of the nuv emission from sne ia , as it suggests that the bright @xmath5 band can be successfully used to measure distances with low intrinsic scatter .
these findings are also interesting for probing the explosion physics of sn ia events , as the dramatic color change with time and the intrinsic variations at the shorter wavelengths provide benchmarks for sn ia modeling . since the b10 and m10 presentations of uvot sn ia photometry , the final photometric reductions for 11 more normal sne ia have been completed .
the photometry for sne 2008hv , 2009ig , and snf2008 - 0514 were presented in brown et al .
( 2012a ) , photometry for 2009dc was presented in silverman et al .
( 2011 ) , and photometry for sn 2011fe was presented in brown et al .
( 2012b ) .
the photometry for sn 2009an , sn 2009cz , sn 2010cr(=ptf10fps ) , ptf09dnl , ptf10icb , sn 2010gn(=ptf10mwb ) , sn 2011by and sn 2011iv is presented in appendix a. all photometry follow the method outlined in brown et al .
( 2009 ) , which uses post - sn template images to remove non - sn counts from the data , and incorporate the updated zeropoints and time - dependent sensitivity corrections of breeveld et al .
results of fitting functions to the data will be presented in a future work , this work will concentrate on color curves .
uvot photometry was supplemented with ground - based photometry for sne 2005df and 2011fe .
australia national university data for 2005df was published in m10 , included because uvot only observed that sn with the uv filters .
stritzinger et al .
( 2010 ) published ubv photometry as part of an optical / nir study of sn 2006dd and three other sne ia in the host galaxy , ngc 1316 .
that data was used , as the uvot optical data at peak suffered severe coincidence loss .
sn 2011fe also suffered severe coincidence loss at optical peak , so @xmath18 and @xmath24 band data was used from richmond & smith ( 2012 ) .
the reddening estimations are shown in table [ extinct ] , and are based upon the sum of the milky way galaxy ( mwg ) reddening and the host reddening .
the mwg reddening was obtained from schlafly & finkbeiner ( 2011 ) via ned and the host reddening was obtained by the phillips peak pseudocolor relation ( phillips 1999 ) , the lira tail color relation ( lira 1995 ) and/or from ground - based studies of these sne ia , the host reddening being the remainder after subtracting mwg reddening . for the majority of this work ,
the reddening estimates are used only for sample selection . in section 5 , the complications of treating extinction in light of newly recognized groupings is discussed .
red - leak corrections , as explained in b10 , were not employed in this work .
k - corrections are estimated using a collection of uv - optical spectra in appendix b , but were not employed for the color curves . the lack of uv spectra to provide adequate time coverage of all the major and minor groups that we have identified in this paper
is the reason that neither correction is implemented for the color curves .
the two filters that suffer the largest effect of red - leak , in terms of the fraction of photons estimated to originate from the wavelength range of the adjacent redder filter , are the @xmath7 and @xmath9 filters . as will be shown ,
the color curves that include these filters are not simply tracings of the color curves from the adjacent redder filter .
this suggests that the most red - leak affected filters still provide unique information about photons in their central wavelength range .
nonetheless , readers are cautioned to avoid treating any of the filters as box functions about a central wavelength .
l|c|ccccc|c sn & @xmath15 & e(b - v)@xmath25 & e(b - v)@xmath26 & e(b - v)@xmath27(tail ) & e(b - v)@xmath27(lit)@xmath28 & ref.@xmath29 & e(b - v)@xmath30(used ) + & [ mag ] & [ mag ] & [ mag ] & [ mag ] & [ mag ] & & [ mag ] + ( 1 ) & ( 2 ) & ( 3 ) & ( 4 ) & ( 5 ) & ( 6 ) & ( 7 ) + + 2005cf & 1.07 & 0.087 & -0.01 @xmath31 0.05 & 0.16 @xmath31 0.04 & 0.11 @xmath31 0.03 & 1,3 & 0.21 + 2005df & 1.20 & 0.026 & 0.13 @xmath31 0.15 & 0.02 @xmath31 0.10 & & 1 & 0.05 + 2006dd & 1.34 & 0.019 & 0.05 @xmath31 0.01 & @xmath31 & 0.04@xmath31 0.01 & 2,9 & 0.06 + 2006 dm & 1.54 & 0.034 & 0.05 @xmath31 0.08 & 0.05 @xmath31 0.01 & 0.04@xmath31 0.09 & 1,3 & 0.09 + 2006ej & 1.39 & 0.031 & 0.18 @xmath31 0.12 & 0.05 @xmath31 0.01 & 0.03 @xmath31 0.09 & 1,3 & 0.09 + 2007af & 1.22 & 0.034 & 0.17 @xmath31 0.08 & 0.05 @xmath31 0.01 & 0.15 @xmath31 0.04 & 1,3 & 0.19 + 2007cq & 1.04 & 0.095 & 0.10 @xmath31 0.12 & & & 1 & 0.21 + 2007cv & 1.31 & 0.062 & 0.14 @xmath31 0.08 & & & 1 & 0.21 + 2007gi & 1.37 & 0.021 & & & 0.16 @xmath31 0.05 & 2,3 & 0.21 + 2007sr & 1.16 & 0.041 & 0.12 @xmath31 0.05 & & 0.11 @xmath31 0.08 & 2,3 & 0.16 + 2008q & 1.40 & 0.073 & 0.02 @xmath31 0.08 & & & 1,8 & 0.10 + snf08 - 0514 & 1.20 & 0.030 & -0.09 @xmath31 0.07 & -0.07 @xmath31 0.16 & & 8 & 0.03 + 2008ec & 1.08 & 0.061 & 0.16 @xmath31 0.08 & 0.20 @xmath31 0.15 & & 1 & 0.25 + 2008hv & 0.95 & 0.029 & -0.12 @xmath31 0.06 & -0.010 @xmath31 0.13 & & 5 & 0.03 + 2009an & 1.20 & 0.016 & & & & 5 & + 2009cz & 0.99 & 0.023 & 0.04 @xmath31 0.01 & & & @xmath32 & 0.07 + 2009dc & 0.72 & 0.062 & -0.29 @xmath31 0.10 & -0.01 @xmath31 0.08 & & 5 & 0.17 + 2009ig & 0.70 & 0.028 & 0.22 @xmath31 0.20 & 0.09 @xmath31 0.06 & 0.00 @xmath31 0.01 & 5,6 & 0.12 + 09dnl & 0.98 & & & & & 10 & + 2010gn@xmath33 & 1.19 & 0.027 & & & 0.05@xmath34 & 10 & 0.05 + 10icb & 1.09 & 0.011 & 0.06 @xmath31 0.04 & & 0.08@xmath34 & 10 & 0.08 + 2011by & 1.14 & 0.012 & & & -0.04@xmath35 & 12 & 0.015 + 2011fe & 1.21 & 0.008 & & & & 7 & 0.025 + + 2010cr@xmath33 & 1.79 & 0.031 & & & & 10 & + 2011iv & 1.69 & 0.010 & & & 0.01 & 11 & 0.01 + [ cols= " < " , ]
@xmath36-corrections are estimated from spectra to account for the effects of redshift on the flux through a filter , as measured here on earth .
although the small aperture of the uvot instrument dictates low - redshift sne ia as targets , the dramatic drop in flux blueward of 4000 typical of sn ia spectra suggests that @xmath36-corrections could even be important in this sample .
the obvious challenge for applying @xmath36-corrections is to accumulate time sequences of uv spectra of all major and minor groups to determine accurate @xmath36-corrections for all sne ia .
recent hst and uvot spectral campaigns are improving the situation ( e.g. foley et al .
2011 , foley et al .
2012 ) , but to date , the spectral sampling of sne ia observed in the uv is : 1 ) not adequate for the generation of @xmath36-correction sequences in the @xmath7 , @xmath9 , @xmath8-bands , 2 ) better in the @xmath6-band and 3 ) unimportant in the @xmath19 and @xmath17-bands . our first effort to generate @xmath36-corrections was presented in b10 , where @xmath36-corrections were obtained by mangling " three spectra ( e.g. a + 5 day spectrum of sn 1992a ( kirshner et al .
1993 ) , a maximum - light template from hsiao et al .
( 2009 ) and a maximum - light template from nugent et al .
( 2002 ) , adjusting the wavelength of each spectrum for the measured redshift of each sn , and then convolving that spectrum with the uvot filter transmissions .
the hsiao template included rest - frame uv spectra from high-@xmath12 sne ia , while the nugent template sequence added @xmath37 spectra of many sne ia to the sn 1992a spectra .
mangling starts with a standard spectrum and distorts it to create spectrophotometry that matches the observed photometry . as a cross - check ,
b10 compared the @xmath36-corrections from the three starting spectra , finding fairly small differences between the three estimates of the @xmath36-corrections .
the sample in b10 included one sn ia that we would now term nuv - blue ( 2008q ) and two that would now be termed muv - blue ( 2007cq & 2006ej ) .
the mangling of the spectra will add emission to account for the nuv - optical excess for the nuv - blue sample , but it is not clear whether the resulting spectral shape is in any way related to the unobserved , true spectral shape . in this appendix , we will investigate only @xmath36-corrections for the @xmath6-band as calculated from the available spectra and from the spectral mangling employed in b10 .
b10 determined that the k@xmath38 and k@xmath39 corrections were typically 0.01@xmath40 0.02@xmath40 , with a maximum of 0.04@xmath40 .
for the current application of color curves and absolute magnitudes , we consider those values negligible , and since the @xmath19 and @xmath17 colors of all major and minor groups are similar , we do not further address k@xmath38 and k@xmath39 .
the @xmath36-corrections reported in b10 were not negligible for the filters bluer than the @xmath19 filter .
the mean corrections reported in b10 are 0.09@xmath40 , 0.15@xmath40 and 0.06@xmath40 for k@xmath41 , k@xmath42 and k@xmath43 , respectively . in order to apply @xmath36-corrections for
the @xmath7 , @xmath9 , @xmath8-bands , the spectra must extend blueward to @xmath101600 and redward to at least 3500 as dictated by the spectral transmission curves of the uvot filters .
few observed uv spectra meet those criteria combined with a uvot color - curve grouping determinations ( as reported in this work ) , meaning that checking the @xmath36-corrections from spectral mangling versus corrections from actual spectra is not possible for the bluest filters . in this current work
, we ignore @xmath36-corrections in the @xmath7 , @xmath9 , @xmath8-bands , rather than employ the corrections from spectral mangling .
this decision was based upon the importance placed on avoiding biasing the presentation of color differences in this initial presentation of these differences .
a future effort will concentrate on trying to derive @xmath36-corrections for all major and minor groups from high-@xmath12 sne ia , where the rest - frame uv is observed in the optical .
that effort hinges on whether there can be unambiguous determination of the sn ia grouping from the available optical observations .
obtaining spectra from low-@xmath12 sne ia that reaches 1800 would also be valuable .
far more spectra span the 3000 4000 wavelength range important for determining k@xmath43 . figure [ kcorru ]
shows k@xmath43 for a range of redshifts as calculated from near optical peak spectra of 5 sne ia .
the nuv - red spectra are a sn 1992a composite spectrum from kirshner et al .
1993 , and two sn 2005cf composite spectra at -0.8@xmath32 and + 2.2@xmath32 , from wang et al . 2011 .
the narrow - peaked spectrum is a composite spectrum of sn 2011iv at + 0.6@xmath32 from foley at al .
the nuv - red - irregular spectrum is the average of two _ swift _ spectra at -4.2@xmath32 and -2.1@xmath32 from foley et al .
the nuv - blue spectrum is a spectrum of sn 2001ay from foley , filippenko & jha 2008 .
the derived k@xmath43 values agree with values from b10 for the majority of sne ia studied .
most importantly , the corrections range from 0.02@xmath40 0.08@xmath40 , suggesting that ignoring the color groupings will induce an error of 0.06@xmath40 or less for sne ia beyond z=0.03 .
there is no clear distinction between the groups , as the nuv - red events have both the largest and smallest corrections in this sample .
the lack of clear distinctions between the groups relative to the variations within each group leads us to not employ k@xmath43 corrections in this work . | observations of many type ia supernovae for multiple epochs per object with the _ swift _ uvot instrument has revealed that there exists order to differences in the uv - optical colors of optically normal supernovae . we examine uv - optical color curves for 23 sne ia , dividing the sne into four groups , finding that roughly one - third of nuv - blue " sne ia have bluer uv - optical colors than the larger nuv - red " group .
two minor groups are recognized , muv - blue " and irregular " sne ia . while the latter group is concluded to be a subset of the nuv - red group , containing the sne with the broadest optical peaks , the muv - blue " group
is concluded to be a distinct group .
separating into the groups and accounting for the time evolution of the uv - optical colors , lowers the scatter in two nuv - optical colors ( e.g. @xmath0 and @xmath1 ) to the level of the scatter in @xmath2 .
this finding is promising for extending the cosmological utilization of sne ia into the nuv .
we generate spectrophotometry of 33 sne ia and determine the correct grouping for each .
we argue that there is a fundamental spectral difference in the 29003500@xmath3 wavelength range , a region suggested to be dominated by absorption from iron - peak elements .
the nuv - blue sne ia feature less absorption than the nuv - red sne ia .
we show that all nuv - blue sne ia in this sample also show evidence of unburned carbon in optical spectra , whereas only one nuv - red sn ia features that absorption line .
every nuv - blue event also exhibits a low gradient of the siii @xmath46355 absorption feature .
many nuv - red events also exhibit a low gradient , perhaps suggestive that nuv - blue events are a subset of the larger lvg group . |
the physical state of the radiation pressure dominated , innermost regions of accretion disks around black holes has been uncertain ever since the early days of accretion disk theory .
standard alpha disk models in which the viscous stress is assumed to scale with the radiation pressure are subject to thermal and viscous instabilities ( lightman & eardley 1974 , shakura & sunyaev 1976 ) .
these instabilities are sensitive to the assumed prescription for the anomalous viscosity ( e.g. piran 1978 ) , so it remains unclear whether or not they are actually present in real flows . if they are , the accretion flow may adopt a radically different state from that usually envisaged in thin accretion disk theory , such as the multiphase equilibrium recently proposed by krolik ( 1998 ) .
in addition to these secular instabilities , dynamical instabilities also exist .
first and foremost , a differentially rotating flow with a negative angular velocity gradient and an initially weak magnetic field is unstable to the magnetorotational instability ( mri , balbus & hawley 1991 ) .
the turbulence resulting from this instability is currently the most plausible mechanism known for generating the anomalous viscosity required in accretion disk models .
dissipation of magnetohydrodynamic waves excited by this turbulence by photon diffusion and photon viscosity has recently been examined by agol & krolik ( 1998 ) .
photon diffusion might also affect the linear development of the instability itself in a laminar , radiation pressure dominated flow , but this issue has not been examined previously . gammie ( 1998 ) has suggested that the overstable photon bubble modes discussed by arons ( 1992 ) in the context of x - ray pulsars also exist in radiation dominated accretion flows in general .
however , his instability analysis was limited to studying a static equilibrium where the effects of differential rotation were entirely neglected .
pietrini & krolik ( 2000 ) have recently investigated convective instabilities in unmagnetized , radiation pressure dominated , differentially rotating flows .
they assumed a constant vertical density profile , which necessarily leads to an unstable entropy gradient . whether or not such unstable gradients exist in reality depends on the vertical dissipation profile , which in turn depends on the structure of the mhd turbulence .
these dynamical instabilities may all play a role at some level in the dynamics and thermodynamics of the radiation pressure dominated portion of accretion disks .
it is the purpose of this paper to provide a unified description of all three by deriving a general linear dispersion relation which incorporates them all .
such an analysis will hopefully prove to be a useful guide to numerical simulations which explore the nonlinear development of these instabilities and their effects on the resulting turbulent state of the inner accretion disk .
this paper is organized as follows . in section 2
we discuss our basic equations and assumptions .
then in section 3 we focus exclusively on the mri by deriving its dispersion relation in the absence of vertical stratification so that the other instabilities are suppressed .
we are therefore able to study just the effects of photon diffusion on what is probably the most important of these instabilities . in section 4
we then present the full dispersion relation for a vertically stratified , differentially rotating medium , and discuss its solutions .
we discuss the relevance of these solutions to astrophysical accretion disks in section 5 and then summarize our conclusions in section 6 .
in eulerian coordinates in an inertial frame , the equations of radiation mhd which we use in this paper are @xmath0 @xmath1 @xmath2 @xmath3 @xmath4 and @xmath5 here @xmath6 is the fluid mass density , @xmath7 is the velocity , @xmath8 is the gas pressure , @xmath9 is the magnetic field , @xmath10 is the gas adiabatic index , @xmath11 is the radiation energy density , @xmath12 is the radiative flux , and @xmath13 is the speed of light .
we have followed the notation of stone , mihalas , & norman ( 1992 ) with regard to variables associated with the radiation field . in order to facilitate comparison with the work of gammie ( 1998 )
, we note that his @xmath14 is our @xmath8 , his @xmath15 is @xmath16 , and his @xmath17 is @xmath18 .
we neglect the self - gravity of the fluid , and assume that the gravitational potential @xmath19 originates from the central mass .
we completely neglect relativistic effects .
we also assume a pure electron scattering opacity @xmath20 , neglecting absorption opacity entirely .
this should be a good first approximation to the inner parts of accretion disks around black holes which are electron scattering dominated .
note , however , that this has the consequence that the gas and radiation are coupled together purely by momentum exchange .
they do not exchange heat , which is why the gas energy equation ( [ dsdt ] ) simply reduces to the condition for adiabatic flow .
this approximation still recovers the basic mri and photon bubble instabilities , and so we believe captures the physics that we wish to explore here . we have closed the radiation moment equations by assuming that the radiation field is close to isotropic , so that the stress tensor is diagonal with elements given by one third the radiation energy density .
we are therefore neglecting the effects of photon viscosity considered for example by agol & krolik ( 1998 ) . the radiation momentum equation ( [ fdiff ] ) deserves special comment .
we have assumed a simple diffusion form , and have in particular neglected time derivative and velocity - dependent terms that , e.g. stone et al .
( 1992 ) and gammie ( 1998 ) have retained .
the full equation is @xmath21 we are interested in perturbations with wavelengths @xmath22 smaller than the vertical disk scale height @xmath23 , and angular frequencies @xmath24 or growth rates which are of order the local angular velocity of the disk @xmath25 ( mri , balbus & hawley 1991 ) or possibly a factor @xmath26 larger ( photon bubbles , gammie 1998 ) .
all the terms on the left hand side of equation ( [ dfdt ] ) will be negligible provided @xmath27 where @xmath28 is the scattering optical depth across a scale height , and @xmath29 is the total ( gas plus radiation pressure ) sound speed . well outside the innermost stable circular orbit , the scattering depth in the inner , radiation pressure dominated zone of a standard @xmath30-viscosity disk ( shakura & sunyaev 1973 ) .
we are nevertheless using @xmath30 to give us some handle on the ambient conditions in a fictitious , laminar accretion disk prior to the development of any instability .
the reader should therefore view these scalings with appropriate caution . ]
is given by @xmath31 here @xmath32 is the radiative efficiency , @xmath33 is the luminosity scaled with eddington , and @xmath34 is the gravitational radius of the central object .
the scattering depth is therefore always significantly greater than unity except perhaps in the innermost parts of disks accreting near the eddington limit .
radiative equilibrium , hydrostatic equilibrium , and the assumption that the viscous stress is @xmath30 times the total pressure leads to a ratio of the speed of light to the sound speed @xmath35 given by @xmath36 hence conditions ( [ dfdtcond ] ) are almost always satisfied in the interior of accretion disks , provided we do not consider wavelengths that are extremely small .
the only exception is for near - eddington disks in the innermost few gravitational radii , where our newtonian treatment of the magnetohydrodynamics breaks down anyway .
we therefore choose to neglect the left hand side of equation ( [ dfdt ] ) , which therefore reduces to the diffusion equation ( [ fdiff ] ) .
the fact that gammie ( 1998 ) found flux limited electromagnetic waves as solutions of his dispersion relation is a reflection of the fact that he retained these terms . in most cases , these terms have negligible effect on the photon bubble modes , which are at much lower frequency . )
included and found no significant differences for realistic physical parameters . ]
note that equation ( [ fdiff ] ) allows us to write the radiation pressure very simply in terms of a gradient in the radiation energy density @xmath11 , so the momentum equation ( [ gasmom ] ) becomes @xmath37 eliminating @xmath12 using equation ( [ fdiff ] ) , we may write the radiation energy equation ( [ dedt ] ) as @xmath38 in contrast to the radiation momentum equation ( [ dfdt ] ) , the left hand side of this equation can be neglected only if @xmath39 because @xmath30 may be quite small and @xmath40 is not very small , these conditions may not be satisfied , so we must retain the left hand side of the radiation energy equation ( [ dedt ] ) .
in fact , it is neglecting the _ right _ hand side of this equation ( photon diffusion ) which returns us to the standard ideal mhd equations with gas and radiation pressure .
when the conditions ( [ diffcond ] ) start to become satisfied , photon diffusion starts to modify standard mhd .
it is this effect which gives rise to photon bubbles , and which can modify the mri .
we therefore choose to retain the left hand side of equation ( [ dedt ] ) , although the right hand side can be dominant .
we consider a narrow annulus of the disk which , in equilibrium , is stationary and axisymmetric with a purely azimuthal flow velocity @xmath41 .
( we assume cylindrical polar coordinates @xmath42 throughout this paper . )
following the usual analysis of the mri as well as previous analyses of photon bubbles , we assume the existence of a locally uniform magnetic field .
because we are considering a differentially rotating system , we assume this field has no radial component , which would otherwise be sheared into a time - dependent azimuthal field .
hence @xmath43 . with the exception of the external gravitational potential @xmath19
, we assume that all radial gradients are much smaller than vertical gradients of the same quantities , i.e. that the disk is geometrically thin .
the radiative diffusion equation ( [ fdiff ] ) then implies that the equilibrium radiative flux is mainly in the vertical direction , and thus the radial component may be neglected .
we will allow for the possibility of radial fluid gradients creating small departures from keplerian flow by retaining a general form for the angular velocity profile @xmath44 in our dispersion relation . with these assumptions , the only nontrivial equilibrium equations are those expressing vertical hydrostatic balance , @xmath45 radiative equilibrium , @xmath46 and radiation diffusion , @xmath47 here @xmath48 is the magnitude of the local vertical gravitational acceleration produced by the tidal field of the central mass .
this equilibrium is characterized by a number of parameters which are important in describing wave modes .
first are the gas and radiation sound speeds , defined as @xmath49 respectively .
then there are the alfvn speeds associated with each component of the magnetic field , @xmath50 internal hydrodynamic modes are driven by equilibrium gradients with associated oscillation frequencies .
the square of the epicyclic frequency is given by @xmath51 it is convenient to define the square of the brunt - visl frequency separately in the gas and radiation as @xmath52\ ] ] and @xmath53,\ ] ] respectively
. then the square of the total brunt - visl frequency is @xmath54={\cgas^2n_{\rm g}^2+\crad^2n_{\rm r}^2\over\cgas^2+\crad^2}. \label{brunt}\ ] ] finally , there is a characteristic length scale associated with radiative diffusion , given by how far photons diffuse on an orbital time .
this can be used to define a diffusion wavenumber as @xmath55 linearizing our equations of motion about the equilibrium , and assuming a wkb space - time dependence @xmath56 $ ] , we arrive at the following perturbation equations : @xmath57 @xmath58 + i{k_{z}\over 4\pi}b_{z}\delta b_{r},\ ] ] @xmath59 @xmath60 @xmath61 @xmath62 @xmath63 @xmath64 @xmath65 @xmath66 @xmath67 and @xmath68 the radial and vertical induction equations together guarantee that the divergence of the magnetic field perturbation vanish .
hence one or the other can be replaced with @xmath69 equations ( [ dcont ] ) and ( [ dengas])-([dfz ] ) can be solved to yield a simple expression for the perturbed total pressure , viz .
@xmath70 where @xmath71 each term on the right hand side of equation ( [ deltaptot ] ) has a simple interpretation . the first is the acoustic pressure perturbation in response to rarefaction or compression .
the second is the buoyancy force ( note that @xmath72 is simply the vertical lagrangian displacement of a fluid element ) .
the last term is the radiation pressure perturbation arising from the enhanced ( diminished ) optical depth in compression ( rarefaction ) regions .
the middle term can give rise to convection , while the last term is responsible for photon bubble modes .
the twelve equations ( [ dcont])-([dbphi ] ) and ( [ divdb ] ) relate twelve perturbation variables .
eight of these equations have an explicit @xmath24 dependence , and so we therefore expect an eighth order dispersion relation .
however , as we show in appendix a , one of these modes is in fact a zero frequency solution which merely represents a perturbation to the equilibrium .
the interesting physics therefore lies in a seventh order dispersion relation . in section 4
, we will present this full dispersion relation , but we will first discuss in section 3 the simpler case of the mri in a non - stratified equilibrium .
in the fictitious laminar flow we are considering , it is presumably the mri which generates the turbulence whose dissipation , combined with cooling , sets up the thermal profile which then might drive photon bubble and convective modes .
we therefore begin by considering the mri in isolation , but in a flow which is still hot enough that radiation pressure dominates gas pressure . in the absence of equilibrium stratification
, the perturbation equations can be solved to yield the following dispersion relation : @xmath73 here @xmath74 @xmath75 are magnetosonic dispersion terms , and @xmath76 is the mri dispersion relation ( if set to zero ) in a non - stratified , incompressible medium ( balbus & hawley 1991 ) .
the dispersion relation ( [ nonstratdisp ] ) describes how the mri is coupled to magnetosonic modes when allowance is made for compressibility and radiation . without radiation ,
it reduces to equation ( 64 ) of blaes & balbus ( 1994 ) for @xmath77 , which is all they considered .
the conclusion of those authors remains valid here : provided the equilibrium azimuthal magnetic field is _ sufficiently _ subthermal , with respect to either the gas or radiation energy densities , then the mri remains essentially unchanged .
mathematically , this is because the alfvn speed is then much smaller than the relevant sound speed . for the mri , @xmath78 , and the dispersion relation reduces to @xmath79 so that the mri mode frequencies remain unchanged from their incompressible form .
physically , in subthermal magnetic fields , the relevant wave speed for the mri ( the alfvn speed ) is much slower than the sound speed , so the associated fluid motions are very nearly incompressible .
the question of how subthermal the azimuthal field must be in order not to affect the mri is determined by radiative diffusion . from the form of the dispersion relation ( [ nonstratdisp ] ) , it is clear that the mri will be unaffected provided @xmath80 a sufficient condition for this to be true is that @xmath81 .
however , in a radiation pressure dominated region of an accretion disk , it is reasonable to consider azimuthal fields for which @xmath82 . in this case
the mri can be modified if the characteristic wavenumber for instability @xmath83 exceeds the diffusion wavenumber @xmath84 , i.e. photons have time to diffuse across a wavelength in an orbital time .
this reduces the effective sound speed , making the fluid more compressible .
the criterion that the mri be unaffected is more stringent : @xmath85 .
compressibility acts to reduce the growth rate of the mri in the presence of azimuthal fields , but it does not remove the instability ( blaes & balbus 1994 ) . ) admits a zero frequency root for @xmath24 when @xmath86 ( see section 4 ) , which is the condition for the marginally stable wavenumber of the mri in incompressible mhd in the absence of stratification . ] this is illustrated in fig .
1 , which depicts the maximum growth rate as a function of azimuthal field strength and diffusion wavenumber .
we have restricted consideration to vertical ( @xmath77 ) modes , as these are expected to have the fastest growth rates , and we consider an equilibrium with negligible gas pressure ( @xmath87 ) . at high values of @xmath88 ,
photon diffusion is negligible , and we recover the mri behavior for an ideal compressible fluid .
subthermal azimuthal fields then produce no effect on the growth rate of the instability , but when the field starts to become thermal in strength , there is a modest reduction in the growth rate . for @xmath89 ,
photon diffusion is important . in this regime
, azimuthal fields can reduce the growth rate quite substantially if @xmath90 .
for the parameters chosen in the figure , this corresponds to @xmath91 for @xmath92 .
higher values of @xmath88 ( less diffusion ) require higher azimuthal field strengths to affect the instability . we can estimate how small the mri growth rate becomes when diffusion is important by setting @xmath93 in equation ( [ nonstratdisp ] ) and then solving it in the limit @xmath94 .
the mri then satisfies @xmath95 for @xmath96 , the unstable growth rate is therefore reduced to @xmath97 .
note that regardless of photon diffusion , the mri is virtually unaffected by compressibility provided the field is sufficiently vertical , i.e. @xmath98 ( e.g. the @xmath99 curve shown in fig .
1 ) . this is because a flow with a purely vertical equilibrium field will have vertical ( @xmath77 ) mri modes identical to the incompressible case , as the dispersion relation ( [ nonstratdisp ] ) permits @xmath100 as a solution . to summarize ,
the mri is essentially unaffected by photon diffusion , unless the azimuthal component of the field is comparable to or larger than the vertical component , and the vertical component of the field is weak enough that photon diffusion is important on the scale of an unstable wavelength @xmath101 .
while the instability still persists , even subthermal azimuthal field components can then reduce the growth rate to @xmath97 if condition ( [ vaphicond ] ) is violated . in the next section we will examine stratified equilibria , which turn out to be subject to a number of unstable modes which can couple to each other in complicated ways . to help disentangle this complexity
, it is helpful to briefly consider the relation between the mri and classical mhd waves .
normally , in the absence of radiative diffusion , the mri can be viewed as a differential rotation driven destabilization of the slow magnetosonic mode in a weakly magnetized medium ( balbus & hawley 1998 ) .
the distinction between slow and alfvn modes is delicate , however . in the limit where the magnetic field becomes very weak
, the slow magnetosonic mode becomes incompressible and has a dispersion relation which is the same as that of alfvn waves .
it is in fact identical in this limit to the `` pseudo - alfvn '' mode which exists in incompressible fluids .
the prefix `` pseudo '' is used because , although the mode is degenerate with the alfvn mode , the eigenfunction is still different .
alfvn waves have velocity perturbations which are perpendicular to the plane of the wave vector and equilibrium magnetic field .
slow , or pseudo - alfvn , waves have velocity perturbations in the plane of the wave vector and equilibrium magnetic field .
for example , if @xmath102 in our geometry , alfvn waves would involve only @xmath103 , whereas slow waves would have nonzero @xmath104 and @xmath105 .
when the coriolis forces associated with rotation are included , radial and azimuthal motions are coupled together , so the distinction between slow and alfvn modes is blurred further .
it is therefore perhaps accidental that the stable modes to which the mri converts are normally the slow modes at high wavenumber .
figure 2(a ) illustrates this fact in the case of negligible radiative diffusion .
however , when radiative diffusion is important , the mri can convert to the alfvn wave [ fig .
2(b ) ] or the slow wave [ fig .
2(c ) ] .
this would only be of academic interest were it not for the fact that it is the slow mode which continuously transforms at low wavenumbers into the unstable photon bubble mode in static , stratified equilibria ( gammie 1998 ) .
the relationship between this instability and the mri in linear theory therefore depends on whether the alfvn or slow mode is more closely tied to the mri .
also , as we will show in the next section , the alfvn mode is itself unstable in stratified equilibria .
with considerable algebra , the general perturbation equations from section 2 can be used to derive the following seventh order dispersion relation : @xmath106\nonumber\\ & & + i2\omega\omega k_zk_r\vaz\vaphi\left[-{k_zk^2f_z\over3\omega_{\rm diff } \rho}-i{k_z^2f_z\over3\omega_{\rm diff}\rho}{1\over\rho } { \partial\rho\over\partial z}-2ik_zg\right]\nonumber\\ & & + k_r^2\tilde\omega^2\left(\cgas^2n_{\rm g}^2+{\omega\over\omega_{\rm diff } } \crad^2n_{\rm r}^2\right ) + { k_r^2k_z\tilde\omega^2f_z\omega^2\over3\omega_{\rm diff}\rho } -{g\over\rho}{\partial\rho\over\partial z}k_r^2\vaphi^2\omega^2=0 . \label{stratdisp}\end{aligned}\ ] ] the first three terms are identical to the non - stratified dispersion relation ( [ nonstratdisp ] ) discussed in the last section .
all other terms arise from the stratification .
note that in order for the wkb approximation to be consistent , we must have @xmath107 , so some of the terms with explicit dependence on the vertical density gradient should probably be taken to be negligible . in a radiation pressure
dominated accretion disk , the wkb approximation also demands that @xmath108 the dispersion relation ( [ stratdisp ] ) reduces to those of previous authors under various limits .
first , in the incompressible limit where @xmath109 , we recover the mri dispersion relation in the presence of vertical stratification ( balbus & hawley 1991 ) , albeit slightly modified by radiative diffusion : @xmath110 \tilde\omega^2 - 4\omega^2k_z^2\vaz^2=0.\ ] ] in addition , if we first assume that gas pressure is negligible ( @xmath111 ) , adopt the ansatz that @xmath112 and then take the short wavelength limit @xmath113 , we obtain two roots given by @xmath114 this is identical to the asymptotic photon bubble instability dispersion relation obtained by gammie ( 1998 , eq .
41 ) , adapted to our geometry . in spite of the presence of rotation , the photon bubble instability survives unscathed in this limit .
finally , if we set the equilibrium magnetic field equal to zero , the equilibrium density to be constant , and the gas pressure to be negligible , then after cancelling two resulting zero modes , the dispersion relation reduces to fifth order : @xmath115 this equation resembles the dispersion relation presented by pietrini & krolik ( 2000 ) when the dissipation terms that they included are neglected .
their dispersion relation still has several additional terms compared to equation ( [ pkdisp ] ) , however .
each of these terms are small and arise from the left hand side of the full radiation momentum equation ( [ dfdt ] ) , which we neglected . despite the complexity of the general dispersion relation ,
the mri turns out to be a robust instability .
an analytic indication that this is so can be found by just setting @xmath116 in equation ( [ stratdisp ] ) , which leads to the following equation : @xmath117- k_r^2k_z^2\vaz^2\cgas^2n_{\rm g}^2=0.\ ] ] if the gas brunt - visl frequency @xmath118 , this equation shows that there will exist a zero frequency root if @xmath119^{1/2}/\vaz$ ] , the usual critical wavenumber for the onset of the mri .
even if @xmath120 , there will be a zero frequency root for @xmath77 and @xmath86 .
while not a rigorous proof , these facts suggest that the critical wavenumber for the onset of the mri is unchanged by the effects of radiation .
hence the instability criterion for the mri is still that the angular velocity decreases outward in magnitude .
these expectations are born out by our numerical calculations .
radiative diffusion does not affect the stability criterion , although it can dramatically affect the growth rates , as discussed in section 3 . in the high wavenumber limit @xmath113 ,
the dispersion relation can be factored to give @xmath121=0.\ ] ] the first factor is a strongly damped radiation diffusion mode , the second describes alfvn waves , and the third corresponds to the fast and slow magnetosonic waves .
rotation is irrelevant at these high wavenumbers .
we have lost the radiation effects on the fast and slow modes because in this limit the wavelength is much less than the diffusion length scale .
it is worthwhile examining the behavior of the mhd wave modes to higher order in @xmath122 , because they all turn out to be unstable for sufficiently high vertical photon fluxes . if , as we have been generally assuming , the radiation and magnetic energy densities in the plasma are much larger than the gas thermal energy density , then the fast and slow modes have frequencies given by @xmath123 \label{fastdisp}\ ] ] and @xmath124 respectively .
the lowest order correction to the alfvn mode is of order @xmath122 , and this correction is real , implying that any instability or damping is at still higher order . as we shall see , instabilities in the alfvn waves peak at finite wavenumbers .
the first imaginary term in equation ( [ fastdisp ] ) represents damping of the fast modes by radiative diffusion .
this term agrees with the optically thin damping rate derived by agol & krolik ( 1998 ) , although the waves need not be optically thin for this rate to apply .
the last term in equation ( [ fastdisp ] ) shows that a sufficiently large vertical radiative flux can overcome the damping and destabilize the fast waves .
the instability criterion can be obtained by comparing the last two terms of equation ( [ fastdisp ] ) : @xmath125 crudely , if the speed at which photons are diffusing upward in the equilibrium exceeds the alfvn speed , then fast modes propagating _ upward _ at some nonzero angle to the vertical will be unstable . for large photon fluxes ,
the growth rate of this instability is @xmath126 the instability criterion is most easily satisfied , and the largest growth rates are achieved , for vertical magnetic fields .
azimuthal fields tend to be stabilizing .
equation ( [ slowdisp ] ) reveals that slow modes propagating _ downward _ at some nonzero angle to the vertical are also unstable at high photon fluxes .
these modes have the fastest growth rates , and at lower values of @xmath127 become the photon bubble modes described by gammie s ( 1998 ) dispersion relation ( [ pbdisp ] ) . adopting gammie s ( 1998 ) photon bubble ansatz of @xmath112 and retaining only the leading order terms as @xmath113 in the general dispersion relation , we find @xmath128 the middle term gives the photon bubble dispersion relation ( [ pbdisp ] ) when use is made of the hydrostatic equilibrium condition ( [ hydrostat ] ) in the radiation dominated limit .
the other terms compete with the photon bubble term at low and high wavenumbers , leading us to expect that the photon bubble dispersion relation ( [ pbdisp ] ) will only hold in the range of wavenumbers given by @xmath129 a condition equivalent to this was also noted by gammie ( 1998 ) .
when the last inequality breaks down at high wavenumbers , the slow mode dispersion relation is better described by equation ( [ slowdisp ] ) .
figure 3 presents a numerical solution to the full dispersion relation ( [ stratdisp ] ) which illustrates these high wavenumber instabilities and compares their growth rates to the analytic expressions above .
in addition to the fast and slow modes , the alfvn modes also appear to go unstable for large photon fluxes .
in contrast to the fast and slow modes , the alfvn wave growth rates tend to zero as @xmath113 .
rotation directly affects the unstable growth rates of the alfvn modes , and also affects the fast and slow modes at low wavenumbers .
we describe this , and the coupling to the mri , in the next subsections .
we now return to low wavenumbers where rotation is important .
we first consider the case where the equilibrium field is vertical ( @xmath102 ) .
the general dispersion relation ( [ stratdisp ] ) then reduces to @xmath130\nonumber\\ & & + k_r^2\tilde\omega^2\left(\cgas^2n_{\rm g}^2+{\omega\over\omega_{\rm diff } } \crad^2n_{\rm r}^2\right ) + { k_r^2k_z\tilde\omega^2f_z\omega^2\over3\omega_{\rm diff}\rho}=0 . \label{vphizero}\end{aligned}\ ] ] note that @xmath100 is a solution of this equation when @xmath77 , so we immediately conclude that in the absence of an azimuthal field , the vertical mri modes are completely unaffected by compressibility , stratification , and radiation .
this is not surprising , as the vertical mri modes involve horizontal motions only , and therefore do not involve any excess forces from the vertical radiative flux . equation ( [ vphizero ] ) is complicated , with many parameters . for numerical computations , we restrict consideration to cases where the gas pressure is negligible , so that @xmath131 and @xmath132 are both zero .
in addition , we neglect convection and couplings to gravity modes by considering only isentropic equilibria , so that @xmath133 . under these assumptions , we may use equation ( [ hydrostat ] ) to write the acceleration due to gravity and the equilibrium vertical density gradient in terms of the photon flux , viz .
@xmath134 hence the only new parameter is @xmath135 .
figure 4 presents numerical calculations of the unstable mode growth rates in the diffusive regime at low wavenumber , for various values of photon flux .
the introduction of stratification causes the slow ( photon bubble ) and alfvn wave instabilities to appear and have increasing growth rate as @xmath135 gets larger .
the fast wave instability appears only after a critical threshold in @xmath135 is reached [ cf . fig .
this is quantitatively consistent with our instability criterion ( [ fastcrit ] ) , which predicts that the fast wave will be unstable for @xmath136 for the parameters chosen in fig .
4(c ) should be compared with fig .
3 which has similar radiation parameters .
note that for the value of @xmath84 chosen in fig .
4 , the mri converts to alfvn waves at @xmath137 [ cf . fig .
2(b ) ] , and indeed the unstable alfvn waves show up for @xmath138 .
the mri itself appears to be largely unaffected by stratification , except that the maximum growth rate increases with increasing @xmath135 .
it is unclear what would happen at higher fluxes because we are already close to the wkb limit ( [ wkbcond ] ) at these high fluxes . given that the vertical ( @xmath77 ) mri modes are unaffected by flux , they are not always the fastest growing modes in the presence of nonzero flux .
modes with a slight nonzero @xmath139 in fact grow faster for the equilibrium parameters shown in fig .
the fact that @xmath135 can affect the mri growth rates for @xmath140 , albeit modestly , suggests that geometrically thick accretion flows with significant radial flux in the equilibrium could have a more marked effect on the mri than the geometrically thin configurations considered in this paper .
the photon bubble mode is always the most rapidly growing of the three mhd wave instabilities that exist in stratified media .
it appears to survive for @xmath141 , albeit at a smaller growth rate than expected from extrapolation of the high @xmath127 behavior .
rotation therefore acts to mildly suppress , and does not kill , this instability when @xmath102 .
we finally turn to the full dispersion relation ( [ stratdisp ] ) with nonzero azimuthal field , but continuing to neglect gas pressure and vertical entropy gradients .
figure 5 presents results for the low wavenumber instability growth rates corresponding the the @xmath102 case shown in fig .
as we stated above , the mri is a robust instability which even in the presence of a nonzero azimuthal field and radiation , still exists for @xmath141 . however , as expected from our results in section 3 , the nonzero azimuthal field produces a marked reduction in the overall growth rate .
the maximum mri growth rates in fig .
5 are comparable to the @xmath77 growth rates shown in fig . 1 at the chosen diffusion wavenumber @xmath92 .
the vertical radiative flux makes little difference to the mri compared to the nonstratified case , just as we found in the previous subsection .
the azimuthal field component has a noticeable effect on the other instabilities .
first , the fast and alfvn waves are quickly stabilized .
this is consistent with our findings in section 4.1 , where we noted that azimuthal fields would stabilize the fast waves . for the parameters chosen in fig .
5 , equation ( [ fastcrit ] ) predicts fast wave stability for @xmath142 .
the photon bubble growth rate also decreases with increasing azimuthal field , in quantitative agreement with equation ( [ pbdisp ] ) . at high values of @xmath143 , this mode appears to split into a high wavenumber part ( the true photon bubble ) , and a low wavenumber part [ cf .
the latter dominates the mri growth rate except at the lowest wavenumbers , at least for the parameters chosen in fig .
5 . at high wavenumbers
, this mode evolves into the alfvn mode .
it turns out that diffusion is actually very important in standard @xmath30-viscosity models of accretion disks around black holes and neutron stars ( shakura & sunyaev 1973 ) . neglecting relativistic correction factors
, one can show that in the radiation pressure dominated inner zone , @xmath144 surprisingly , this relation is independent of accretion rate , radius , and central object mass . combining this with the wkb condition for the mri , @xmath145 , we find @xmath146 depending on the assumed strength of the initial magnetic field , realistic accretion disk models ( @xmath147 ) will often be in the regime where photon diffusion can significantly slow the growth of the mri in the presence of azimuthal field components ( cf .
our analysis shows that in this case the growth rate would be reduced to @xmath97 . in the radiation pressure dominated inner zone
, one can conceive of @xmath148 being somewhat less than @xmath149 . neglecting relativistic correction factors once again , @xmath150 where @xmath151 is the mass of the central object .
the mri growth rate can therefore become quite sluggish if the azimuthal field energy density exceeds the thermal energy density in the gas by an amount comparable to the radiation / gas thermal energy density ratio .
we stress however that even in this regime , the mri still exists as a dynamical instability , and will still presumably grow eventually to nonlinear amplitude
. it might be interesting to see if the resulting turbulence depends on how the initial magnetic field energy density compares to the gas and radiation thermal energy densities , but it is difficult to predict any results from linear theory alone .
note that all these conclusions refer to a fictitious , unstable initial state which can not exist in nature .
the equations above , in particular that for the diffusion wavenumber ( [ eqkdiff ] ) , depend on @xmath30 , which is presumably a consequence of the mri .
it is therefore difficult to determine how the mri will be affected when it is responsible for the ambient conditions in the disk in the first place .
this fact is even more important when we consider the stratified calculations that we presented in section 4 .
multiple instabilities are simultaneously present in this case , and the modes with the fastest growth rates presumably dominate the nonlinear evolution .
one would at first conclude that the photon bubble / slow mode at high wavenumbers therefore controls the dynamics . from equation ( [ slowdisp ] ) , the growth rate is at most @xmath152 substantially faster than the orbital frequency @xmath25 . however , it is important to recognize the different sources of free energy which are driving these instabilities .
the mri feeds off the differential rotation in the flow , transporting angular momentum outward which ultimately allows material to accrete and release gravitational potential energy .
we therefore view this instability as being fundamental to the dynamics of the accretion disk .
dissipation of the gravitational energy released and radiative transport processes set up vertical gradients which drive the photon bubble and other mhd wave instabilities .
convection will also presumably occur if the resulting entropy gradient has the right sign ( pietrini & krolik 2000 ) .
it is the vertical gradients which provide the source of free energy for these extra instabilities , but these gradients are fundamentally determined by the action of the mri .
we speculate that on the largest vertical scales , where diffusion is negligible on an orbital time , the mri will always operate as usual , launching perhaps a turbulent cascade to smaller scales .
how this energy is ultimately dissipated is still unclear
. if there is sufficient vertical flux , it may not be as simple as radiative damping of compressive mhd waves ( agol & krolik 1998 ) , as those waves are themselves unstable for some propagation directions .
given the high growth rates at short wavelengths from equation ( [ pbhighk ] ) above , it seems likely to us that photon bubbles will play an important role in thermal transport on small scales .
concrete answers to the question of how radiative diffusion and radiative flux really affect the turbulent angular momentum and heat transport in the disk will require investigation by numerical simulations .
one must be careful to choose the fictitious unstable initial state with care when pursuing such simulations , keeping in mind that accretion power , and therefore the mri , is the driver for all this complex dynamics .
we may summarize our conclusions as follows . in the absence of an azimuthal field ,
the vertical ( @xmath77 ) mri modes have the standard , incompressible behavior elucidated by balbus & hawley ( 1991 ) .
if the azimuthal field component dominates the vertical field component and the alfvn speed is larger than the effective sound speed [ cf .
( [ vaphicond ] ) ] , then the growth rate of the mri is reduced to @xmath97 , well below the usual rapid growth rate of order the orbital frequency @xmath25 . however , the instability still persists .
large vertical photon fluxes in the equilibrium do not significantly affect the most rapidly growing mri modes . at short wavelengths where rotation is unimportant ,
both the fast and slow mhd waves are destabilized by sufficiently high vertical photon flux in the equilibrium .
these instabilities require that the waves propagate at some nonzero angle to the vertical , and affects upward propagating fast waves and downward propagating slow waves .
the slow waves appear to always have the larger growth rates . at smaller wavenumbers these
are well - described by gammie s ( 1998 ) photon bubble dispersion relation ( [ pbdisp ] ) , but at larger wavenumbers they asymptote to a maximum value given by equation ( [ slowdisp ] ) . at small wavenumbers where rotation is important
( @xmath141 ) , the photon bubble growth rate is somewhat suppressed by rotation , at least at modest fluxes .
nonzero azimuthal fields stabilize the fast wave instability and reduce the photon bubble growth rate .
in addition to the fast and slow wave instabilities , the vertical radiative flux also drives unstable modes at intermediate wavenumber which are related to alfvn waves at high wavenumber .
the ambient conditions in the radiation pressure dominated inner region of standard alpha disk models are generally in a regime where radiative diffusion can slow the growth of the mri . in addition
, photon bubble modes are expected at short wavelengths with growth rates that can significantly exceed the orbital frequency .
it is not clear how these instabilities will evolve into the nonlinear , turbulent regime , but we speculate that the mri will dominate the largest scales and that photon bubbles ( and perhaps unstable fast and alfvn waves ) may affect both the dissipation and thermal transport on smaller scales .
a complete understanding of the turbulent angular momentum and heat transport will require numerical simulations .
finally , we stress that we have concentrated here on geometrically thin accretion disk configurations where the photon flux is primarily in the vertical direction and the rotation curve is near - keplerian . geometrically thick or quasi - spherical ,
radiation pressure supported flows will have radiation pressure gradient components in the radial direction and may therefore exhibit more interesting dynamical coupling between the radiative flux and the mri .
despite the fact that the perturbation equations ( [ dcont])-([dbphi ] ) and ( [ divdb ] ) have eight separate @xmath24-dependent terms , we only end up with a dispersion relation that is seventh order .
the reason is that there is a zero frequency mode , which we discuss in some detail in this appendix . setting @xmath116 in the perturbation equations ,
the radial induction equation ( [ dbr ] ) immediately implies that @xmath104 must vanish . the continuity equation ( [ dcont ] ) and azimuthal momentum equation ( [ dvphi ] )
then imply that @xmath105 and @xmath153 must also vanish .
the gas energy equation ( [ dengas ] ) is then trivially satisfied , which is the mathematical reason why a zero frequency mode is a possible solution of the perturbation equations .
excluding the trivial azimuthal flux equation ( [ dfphi ] ) , the remaining seven equations may be written @xmath154 @xmath155 @xmath156 @xmath157 @xmath158 @xmath159 and @xmath160 these seven equations have eight unknowns , so that any one of the nonzero perturbation amplitudes determine the other seven for this zero frequency mode .
it is perhaps surprising that this mode has such a complicated eigenfunction , and it is therefore worthwhile understanding its physical meaning .
equation ( [ zshear ] ) is particularly interesting .
this arose from the azimuthal component of the induction equation which describes , for example , how a radial magnetic field component is sheared out by the differential rotation to produce a time - dependent azimuthal field component . in this case , however , this is exactly compensated by an azimuthal velocity perturbation which creates azimuthal field in the opposite direction out of vertical field .
further insight into the physical nature of this mode may be obtained by taking various limits . in the limit @xmath161 , we have @xmath162 , and the only nontrivial condition comes from the radial momentum equation : @xmath163 in other words , any random field of perturbations of @xmath103 can be balanced by radial gradients of gas and/or magnetic pressure to produce a slightly different equilibrium .
in the opposite limit @xmath164 , we have @xmath165 , and @xmath155 @xmath157 @xmath166 in this limit , a random field of density perturbations is balanced by vertical gradients in gas and radiation pressure , again producing a slightly different equilibrium . finally , we note that if we neglect the vertical stratification of the equilibrium , then this zero mode has all perturbations vanishing except for the density . in this case a random field of density perturbations will merely be advected along with the equilibrium flow without producing additional net accelerations . | we present a general linear dispersion relation which describes the coupled behavior of magnetorotational , photon bubble , and convective instabilities in weakly magnetized , differentially rotating accretion disks .
we presume the accretion disks to be geometrically thin and supported vertically by radiation pressure .
we fully incorporate the effects of a nonzero radiative diffusion length on the linear modes . in an equilibrium with purely vertical magnetic field ,
the vertical magnetorotational modes are completely unaffected by compressibility , stratification , and radiative diffusion .
however , in the presence of azimuthal fields , which are expected in differentially rotating flows , the growth rate of all magnetorotational modes can be reduced substantially below the orbital frequency .
this occurs if diffusion destroys radiation sound waves on the length scale of the instability , and the magnetic energy density of the azimuthal component exceeds the non - radiative thermal energy density . while sluggish in this case , the magnetorotational instability still persists and will still tap the free energy of the differential rotation .
photon bubble instabilities are generically present in radiation pressure dominated flows where diffusion is present .
we show that their growth rates are limited to a maximum value which is reached at short wavelengths where the modes may be viewed as unstable slow magnetosonic waves .
we also find that vertical radiation pressure destabilizes upward propagating fast waves , and that alfvn waves can be unstable .
these instabilities typically have smaller growth rates than the photon bubble / slow modes .
we discuss how all these modes behave in various regimes of interest , and speculate how they may affect the dynamics of real accretion disk flows .
# 1to 0pt#1 = cmbsy10 = cmmib10 |
recently , there has been much interest on two - way relay networks ( twrns ) in which two source nodes @xmath0 and @xmath1 without a direct link communicate with each other via a relay node .
the architecture of twrns makes it possible to better exploit the channel multiplexing of uplink and downlink wireless medium @xcite .
the source nodes initially send their data to the relay node .
the received data is combined employing a certain method according to the amplify - and - forward ( af ) or the decode - and - forward ( df ) mode and gets broadcasted from the relay back to both source nodes . with the application of network coding and channel estimation techniques
@xcite , @xmath0 and @xmath1 can perform self - interference cancelation and remove their own transmitted codewords from the received signal .
four time slots needed in a traditional one - way transmission for the forward and backward channels to accomplish one - round information exchange between @xmath0 and @xmath1 via the relay node can be reduced to two in twrns by comparison . in a realistic multi - user wireless network , e.g. is-856 system @xcite which has more relaxed delay requirements , the transmission power is fixed while the rate can be adapted according to the channel conditions . moreover , automatic repeat request ( arq ) techniques have been applied to improve the transmission reliability above the physical layer @xcite . prior works @xcite , @xcite also show there is a compromise between transmission rate @xmath2 and arq such that the network average successful throughput , i.e. , the goodput , can be maximized at an optimal rate @xmath3 .
in addition to the goodput analysis , we in this paper are interested in energy - efficient operation . in such cases , the energy consumption due to retransmission should also be taken into account to evaluate the energy efficiency with respect to @xmath2 . hence , we investigate the joint optimization of @xmath2 ( at the physical layer ) and the number of arq retransmissions ( at the data - link layer ) by adopting a cross - layer framework in twrns . the remainder of this paper is organized as follows .
section ii introduces the twrn model , channel assumptions as well as its general working mechanism . in sections
iii and iv , we investigate the markov chain model under both af and df modes to derive the analytical expressions for the bit energy consumption and goodput . in section
v , the numerical results are shown to compare the system performance in af and df modes .
section vi provides the conclusion .
in figure 1 , we depict a 3-node twrn where source nodes @xmath0 and @xmath1 can only exchange information via the relay node . codewords @xmath4 and @xmath5 from @xmath0 and @xmath1 , respectively , have equal length and unit energy .
all nodes are working in half - duplex mode and the channels between @xmath0 and the relay , and @xmath1 and the relay are modeled as complex gaussian random variables with distributions @xmath6 and @xmath7 . without loss of generality , we also assume channel reciprocity such that @xmath8 and @xmath9 have identical distributions as @xmath10 and @xmath11 , respectively .
odd and even time slots have equal length , which is the time to transmit one codeword , and are dedicated for uplink and downlink data transmissions , respectively .
ack and nack control packets are assumed to be always successfully received and the trivial processing time is ignored .
additive gaussian noise at the receiver terminals is modeled as @xmath12 .
there are two more key assumptions : 1 ) channel codes support communication at the instantaneous channel capacity levels , and outages , which occur if transmission rate exceeds the instantaneous channel capacity , lead to packet errors and are perfectly detected at the receivers ; 2 ) depending on whether packets are successfully received or not , ack or nack control frames are sent and received with no errors . based on above network formulations
, we can further discuss the twrn working procedure according to the current network states under af and df relay schemes , and find out the inherent impact of the transmission rate @xmath2 on network performances .
the twrn in af mode can be visualized as two bi - directional cascade channels where in the odd time slots , @xmath0 and @xmath1 send individual codewords simultaneously to the relay and the signals are actually superimposed in the wireless medium .
the relay will then amplify the received signals proportional to the average received power and broadcast the combined signals back to @xmath0 and @xmath1 in the even time slots . according to fig.[fig : twrn ] , the received signals at the relay in odd time slots is @xmath13 where @xmath14 , @xmath15 are the transmit power of @xmath0 and @xmath1 respectively
. the relay will forward @xmath16 with a scaling factor @xmath17 which is @xmath18 where @xmath19 is the relay s transmit power . here
, we normalize the variance of the channel between @xmath0 and @xmath1 as @xmath20 and by using a normalized distance factor @xcite @xmath21 while assuming the variances of the other links are proportional to @xmath22 where @xmath23 is the path loss coefficient , we then have @xmath24 and @xmath25 . at the end of even time slots , the received signals on @xmath0 and @xmath1 can be written as @xmath26 where @xmath27 , @xmath28 , and @xmath29 are i.i.d gaussian noise components . assuming the instantaneous channel state information is perfectly known at @xmath0 and @xmath1 , the self - interference part can be removed from @xmath30 and @xmath31 and the signals for decoding can be represented by @xmath32 the cascade channel instantaneous rate from @xmath0 to @xmath1 and from @xmath1 to @xmath0 are hence represented by @xmath33 to describe the network mechanism more accurately , we need to formulate the protocol of twrn in af mode as follows : 1 .
each transmission round contains two consecutive time slots . in the odd slot , source nodes @xmath0 and @xmath1
both transmit codewords to the relay with transmission rate @xmath2 bit / sec / hz , and the relay in the following even slot broadcasts @xmath34 back to source nodes .
2 . at the end of one transmission round , @xmath0 and
@xmath1 perform self - interference cancelation ( sic ) to subtract their own weighted messages , and decode @xmath35 and @xmath36 , respectively .
if the decoding fails , an outage event will be declared on that cascade link .
the outage event on the cascade link @xmath37 ( or @xmath38 ) is defined as the probability of the event @xmath39 ( or @xmath40 ) .
ack or nack packets would be sent back to the relay based on successful transmission or outage .
the relay will also notify @xmath0 and @xmath1 whether a new codeword or an old codeword should be ( re)transmitted in the next odd time slot with the control packet information .
the network state transition diagram of the af twrn can be modeled as a markov chain as shown in fig .
[ fig : af - markov ] , where the probability on each path denotes the probability of the transition between two states . @xmath41 and @xmath42 are defined as the outage probabilities on the cascade @xmath37 and @xmath38 links and are given by @xmath43 ( [ eq : af - outage0 ] ) and ( [ eq : af - outage ] ) can be determined using the cumulative distribution function of the random variable @xmath44 @xcite @xmath45 and @xmath46 and @xmath47 are independent exponential distributed with mean @xmath48 and @xmath49 , and @xmath50 is a constant . in this context , we know @xmath51,\,\hspace{0.3cm}\mu_{2}=e[|h_{r2}|^2 ] , \hspace{0.3 cm } for \hspace{0.3 cm } p_{12}\\ \mu_{1}=e[|h_{2r}|^2],\,\hspace{0.3cm}\mu_{1}=e[|h_{r1}|^2 ] , \hspace{0.3 cm } for \hspace{0.3 cm } p_{21}\\ \alpha=\frac{1}{\beta^2 } \end{array } \right.\end{aligned}\ ] ] with the given network parameters , @xmath41 and @xmath42 can be derived accordingly . in fig .
[ fig : af - state ] , it is explicitly seen at the beginning of each odd time slot that both @xmath52 and @xmath53 transmit to the relay such that at the beginning of each even time slot , the relay is always in the ready - for - broadcasting state @xmath54 and will consequently transition to @xmath55 , @xmath56 , @xmath57 , or @xmath58 with certain probabilities .
we first determine the following four equations according to the state transitions in the markov chain to derive the probability of each state : @xmath59 after solving the set of equations in ( [ eq : af - markov - equations ] ) , we obtain @xmath60 . we know from the inherent characteristics of the af twrn that the data exchange only happens in the broadcasting phases with successful packet delivery . therefore , the system goodput is defined similarly as in @xcite by @xmath61 ( [ eq : af - goodput ] ) indicates through the terms @xmath41 and @xmath42 that a higher transmission rate @xmath2 will result in higher packet error rates ( outage ) , leading to more arq retransmissions which equivalently reduce the data rate .
intuitively , a balance between transmission rate and the number of arq retransmissions needs to be found such that the goodput is maximized .
energy efficiency has always been a major concern in wireless networks .
recently , power or energy efficiency in wireless one - way relay networks have been extensively studied . in @xcite ,
the average bit energy consumption @xmath62 is minimized by determining the optimal number of bits per symbol , i.e. , the constellation size , in a specific modulation format .
similarly as in one - way relay channels , the outage probabilities in twrns are functions of the transmission rate @xmath2 .
for instance , there could be an increased number of outage events on the cascade channels when codewords are transmitted at a high rate .
in such a case , more retransmissions and higher energy expenditure are needed to accomplish the reliable packet delivery . therefore , we are interested in a possible realization of twrn operation , which can provide a well - balanced performance on both the goodput @xmath63 and the required energy .
we evaluate this by formulating the average bit energy consumption @xmath62 required for successfully exchanging one information bit between @xmath0 and @xmath1 .
we then evaluate @xmath62 by considering long - term transmissions on twrn .
regardless of the previous state , whenever the relay is in state @xmath54 and is broadcasting , the resulting state would be any of the other four states previously described .
assuming there are @xmath64 rounds of two - way transmission , each of which consists of a pair of consecutive time slots and each codeword has @xmath65 bits .
therefore , with @xmath66 , where @xmath67 is the number of transmission round corresponding to state @xmath68 , the average bit energy consumption could be derived as the ratio of total bits successfully exchanged over total energy consumption : @xmath69
the df twrn differs from the af twrn in that there is a crucial intermediate decoding procedure at the the relay when it has received the codeword from the uplink transmission .
if both source nodes are allowed to send codewords to the relay simultaneously in the uplink transmission , the decoding at the relay has to deal with the multiple access problem in a realistic application , with successive interference cancelation techniques . to reduce the hardware complexity and increase the feasibility of implementation
, we hereby adopt the df twrn mode from @xcite where the relay performs sequential decode - and - forward .
the outage probabilities on @xmath70 , @xmath71 , @xmath72 , @xmath73 links are denoted as @xmath74 , @xmath75 , @xmath76 and @xmath77 . the protocol for sequential df twrn
is described as follows : 1 . in the initial state @xmath55 ,
the relay s buffer is empty and the relay first polls on @xmath0 until it receives codeword @xmath78 successfully with probability @xmath79 .
then , the state moves to @xmath56 which means the relay holds @xmath4 in the buffer .
otherwise , the state remains as @xmath55 with probability @xmath74 .
if the relay already has @xmath4 , it starts polling @xmath1 .
the state @xmath56 either changes to @xmath58 with probability @xmath80 upon successfully receiving @xmath5 , or stays in @xmath56 with @xmath75 .
3 . when the relay has both @xmath4 and @xmath5 , it generates a new codeword @xmath81 according to a gaussian codebook of @xmath82 with equal power allocation .
then at rate @xmath2 , it broadcasts to @xmath0 and @xmath1 , which will perform sic to decode @xmath5 and @xmath4 , respectively .
accordingly , the state will transit to @xmath55 , @xmath56 , @xmath57 or @xmath58 with corresponding probabilities @xmath83 , @xmath84 , @xmath85 or @xmath86 respectively .
4 . at the beginning of next transmission round
, the relay will decide to poll a new codeword from @xmath0 ( or @xmath1 ) based on the previous state being @xmath55,@xmath57 ( or @xmath56 ) or just retransmits the old @xmath87 if the previous state was @xmath58 .
the network state transition diagram of the df twrn can be modeled as in fig .
[ fig : df - markov ] with detailed probabilities on each path .
since the relay receives and decodes @xmath4 and @xmath5 at different time slots , the received signals at the relay from uplink transmissions can be represented as @xmath88 similar to ( [ eq : af - sic ] ) , the signals for decoding at @xmath0 and @xmath1 after sic has been performed can be written as @xmath89 similarly as in the discussion of the goodput of af twrn in section iii , the data exchange only occurs upon the successful signal receptions at @xmath0 and @xmath1 at the end of the broadcasting time slot .
hence , initially , it is necessary to calculate the probability of being in state stays 3 , i.e. , @xmath90 .
we start from calculating the outage probabilities on the forward and backward channels as @xmath91 the probabilities of buffer states can be solved by noting the following relations from fig .
[ fig : df - markov ] : @xmath92 solving the equations in ( [ eq : df - markov - equations ] ) with given outage probabilities , we can obtain the following results for the buffer states : @xmath93 where the polynomial in the denominators is denoted by @xmath94 therefore , the system goodput in the df mode can be derived as @xmath95 @xmath62 in the df twrn is more complicated to calculate than in the af twrn where each transmission round has fixed power as can be seen in ( [ eq : af - eb ] ) .
hence , in the df scenario , we have to separate the energy expenditure into two parts , energy consumption in the first stage and energy consumption in the second stage .
the first stage denotes the state transition from any of 4 previous states to state @xmath58 , where the relay holds two codewords @xmath4 and @xmath5 in its buffer and is ready to broadcast .
the second stage is that the relay broadcasts its newly generated codeword and the state transits back to any of the four states again .
considering the relay s buffer is to be loaded with both codewords @xmath4 and @xmath5 from any of the previous states on the first stage , the energy consumption conditioned on the previous state @xmath55 , @xmath56 , @xmath57 , or @xmath58 on this particular transition will be @xmath96 on the second stage , the energy consumption for broadcasting is always @xmath97 , so the average bit energy consumption for one information bit successfully exchanged on the df twrn can be computed as @xmath98 whenever the state probabilities and outage probabilities are known .
in this section , we present the numerical results to evaluate the system performance of twrn in both af and df modes .
the network configurations are assumed to be as follows : relay is located in the middle between @xmath0 and @xmath1 which means @xmath99 .
the power spectrum density of the gaussian white noise is @xmath100 and the channel bandwidth is set to @xmath101 hz .
path loss coefficient is @xmath102 @xcite .
we also assume the same transmit power for both source nodes and the relay , which is @xmath103 and define the snr by @xmath104 .
firstly , we are particularly interested in how the goodput varies as a function of the transmission rate @xmath2 at specific snr values . in fig .
[ fig : eff - r ] , @xmath105 and @xmath106 are plotted as functions of @xmath2 , with solid and dashed lines corresponding to af and df modes , respectively . on each curve with a given specific @xmath107 value ,
it s immediately seen that the goodput first increases within low @xmath2 range and then begins to drop once the rate is increased beyond the optimal @xmath3 which maximizes the goodput @xmath63 .
additionally , at low values of the rate @xmath2 , the af twrn has higher goodput @xmath105 , while beyond a certain rate @xmath2 , df starts to outperform , regardless of the snr @xmath107 value . to better illustrate the goodput performance in af and df modes , we look into transmission efficiency by defining a normalized rate @xmath108 and plotting it in fig .
[ fig : normal - r ] .
the normalized rate is always decreasing when @xmath2 increases at all snr s in both af and df .
in other words , increasing outage probabilities due to increasing @xmath2 has eventually resulted in more arq retransmissions . specifically in the high snr scenario ,
the normalized rate levels off between 0.6 and 0.7 in af and 0.9 and 1 in df , which means the transmission efficiency does nt change too much within this rate range .
in addition , the af mode seems to have higher normalized rate in low rate range . in fig .
[ fig : eb - r ] , we analyze the energy efficiency .
we notice that the difference of the average bit energy consumptions between two modes is insignificant up until @xmath109 , but df stills has a better energy efficiency with a lower @xmath62 regardless of @xmath107 .
however when compared with the corresponding points ( @xmath110 ) in fig .
[ fig : normal - r ] , it is shown that even though df can achieve a slightly lower @xmath62 than af , it also suffers a lower transmission efficiency in the metric of lower normalized rate . above @xmath109 in low snr scenario of @xmath111 , af predominates with both higher normalized rate and lower @xmath112 until @xmath2 approaches about @xmath113 bits / sec / hz .
similar results can be observed on the high snr scenarios also . in fig .
[ fig : normal - snr ] , we study the impact of snr on the normalized rate in both af and df .
basically the normalized rate is increasing as snr increases at all transmission rates . at low snr e.g. @xmath114 , df performs better than af .
as snr approaches @xmath115 , the normalized rate gets close to 1 ( or 0.7 ) in af ( or df ) mode .
consequently , one way to improve on the transmission efficiency is to increase snr in the twrn .
considering overall impacts of rate @xmath2 and snr , we can always find an scheme for the twrn to achieve optimality in respect to the goodput @xmath63 , the average bit energy consumption @xmath62 or the transmission efficiency .
in this paper , we have studied the two - way relay networks working in amplify - and - forward and decode - and - forward modes . in each mode
, we set up a markov chain model to analyze the state transition in details .
arq transmission is employed to guarantee the successful packet delivery at the end and mathematical expressions for the goodput and bit energy consumption have been derived .
several interesting results are observed from simulation results : 1 ) the transmission rate @xmath2 can be optimized to achieve a maximal goodput in both af and df modes ; 2 ) generally the transmission efficiency is higher in af within a certain @xmath2 range , while the df can achieve a slightly higher energy efficiency instead ; 3 ) increasing snr will always increase the normalized rate regardless of @xmath2 .
hence , it s possible the network performance be optimized in a balanced manner to maintain a relatively high goodput as well as a low @xmath62 .
v. erceg , l. j. greenstein , s. y. tjandra , s. r. parkoff , a. gupta , b. kulic , a. a. julius , and r. bianchi , `` an emperically based path loss model for wireless channels in suburban environments , '' ieee _ j.select .
areas commun , _ vol .
17 , no . 7 , pp.1205 - 1211 , jul . 1999 . | in this paper , we study two - way relay networks ( twrns ) in which two source nodes exchange their information via a relay node indirectly in rayleigh fading channels .
both amplify - and - forward ( af ) and decode - and - forward ( df ) techniques have been analyzed in the twrn employing a markov chain model through which the network operation is described and investigated in depth .
automatic repeat - request ( arq ) retransmission has been applied to guarantee the successful packet delivery .
the bit energy consumption and goodput expressions have been derived as functions of transmission rate in a given af or df twrn .
numerical results are used to identify the optimal transmission rates where the bit energy consumption is minimized or the goodput is maximized .
the network performances are compared in terms of energy and transmission efficiency in af and df modes . |
the notion of a collection of containers for a hypergraph was introduced by the authors in @xcite .
a collection of containers for a hypergraph @xmath1 is a collection @xmath2 of subsets of @xmath3 such that every independent set @xmath4 is a subset of some member @xmath5 .
( a subset of @xmath3 is independent if it contains no edge . )
the notion was developed further in @xcite and several applications given ; related methods and results were proved by balogh , morris and samotij @xcite .
these results have since been applied by other authors .
our purpose here is to revisit the method of @xcite , and to combine it with a twist that makes it much more widely applicable .
it is true that the method of @xcite is not too complicated , and the consequences are often best possible , but it is subtle .
the method of @xcite , on the other hand , is not optimal ; nevertheless it is very simple , and it is particularly transparent .
it is sufficient , for example , for counting the number of @xmath0-free hypergraphs ( see corollary [ cor : ffree_count ] ) , and hence it offers a very elementary and straightforward proof of this result .
the method of @xcite applies to _ simple _ or _ linear _ hypergraphs , that is , hypergraphs in which no two edges share more than one vertex .
the container theorem there was as follows .
we use the term _
@xmath6-graph _ to mean an @xmath6-uniform hypergraph , where @xmath7 always .
[ prop : curly_c ] let @xmath1 be a @xmath8-regular simple @xmath6-graph .
if @xmath8 is large , there is a collection of sets @xmath9 of subsets of @xmath3 satisfying * if @xmath10 is independent , there is some @xmath11 with @xmath12 , * @xmath13 for every @xmath14 , * @xmath15 where @xmath16 .
this proposition is not quite as stated in @xcite , but it is pretty much explicit in the proof of theorem 1.1 that follows theorem 3.1 .
two drawbacks limit the applicability of proposition [ prop : curly_c ] .
the first is that many popular container applications require containers with @xmath17)$ ] small , rather than @xmath18 small .
the second is that it applies only to regular @xmath6-graphs .
container results are more useful when they can be applied iteratively .
that is , given an independent set @xmath4 in a hypergraph @xmath1 , we can apply the proposition once to obtain a container @xmath19 for @xmath4 , but we would then like to apply the proposition again , this time to the hypergraph @xmath20 $ ] ( of which @xmath4 is still an independent set ) , thus finding a smaller container @xmath21 .
if possible we would then repeat this procedure until very small containers are obtained . the snag with this procedure as it stands
is that it is unlikely that @xmath20 $ ] is regular even if @xmath1 is , and so iteration is not possible .
of course , proposition [ prop : curly_c ] still applies to graphs that are `` somewhat '' regular ( indeed , this follows directly from theorem 3.1 of @xcite ) , but not in a particularly strong way .
both these drawbacks can be overcome by adapting the proof of theorem 3.1 in @xcite to use the notion of _ degree measure _
, which we discuss in [ sec : degree ] .
this yields a version of proposition [ prop : curly_c ] in which @xmath19 is bounded in degree measure , namely theorem [ thm : simplecont ] .
iterated applications of this theorem give the following result , much stronger and more useful than proposition [ prop : curly_c ] , and the main theorem of the present paper . [ thm : sparsecont ]
let @xmath1 be a simple @xmath6-graph of average degree @xmath8 .
let @xmath22 .
if @xmath8 is large enough , then there is a collection of sets @xmath9 of subsets of @xmath3 satisfying * if @xmath10 is independent , there is some @xmath11 with @xmath12 , * @xmath17)<\delta e(g)$ ] for every @xmath14 , * @xmath23 where @xmath24 .
observe that theorem [ thm : sparsecont ] differs from proposition [ prop : curly_c ] only in that the condition of @xmath8-regularity is replaced by that of average degree @xmath8 , and the conclusion giving a bound on @xmath18 is replaced by a bound on @xmath17)$ ] .
this bound on @xmath17)$ ] implies , for regular @xmath1 , that @xmath25 ( see the discussion in [ sec : degree ] ) , which is essentially best possible , but is nonetheless a weaker condition than the bound on the number of edges . thus theorem
[ thm : sparsecont ] is a generalization of proposition [ prop : curly_c ] .
the notion of degree measure was introduced in @xcite . in the definition
below , @xmath26 is the degree of the vertex @xmath27 . [ def : mu ] let @xmath1 be an @xmath6-graph of order @xmath28 and average degree @xmath8 .
let @xmath29 .
the _ degree measure _ @xmath30 of @xmath31 is defined by @xmath32 we note some immediate properties of degree measure .
first , for any @xmath29 , @xmath33)\,\le\ , { 1\over r}\sum_{v\in s } d(v)\ , = \ , { d|g|\over r}\mu(s)\,=\,\mu(s)e(g)\,,\ ] ] so @xmath30 small implies @xmath34)$ ] small .
moreover , sets of large measure must contain many edges .
indeed , writing @xmath35 for @xmath36 and @xmath37 for the number of edges meeting both @xmath35 and @xmath31 , we have @xmath38)\right\ } \,,\ ] ] that is , @xmath39)/nd$ ] .
since @xmath40 this means @xmath41)\,\ge\,(\mu(s)-1+{1\over r})\,nd\,.\ ] ] in particular , the measure of an independent set can not exceed @xmath42 .
furthermore , if @xmath1 is regular , then degree and uniform measures coincide ; in this case , the inequality @xmath34)\le
\delta e(g)$ ] together with ( [ eqn : mubig ] ) implies @xmath43 , as mentioned in the introduction .
we can now state the theorem which is at the heart of the present paper .
this theorem is already sufficiently powerful for obtaining non - trivial results , such as in list colouring .
[ thm : simplecont ] let @xmath1 be a simple @xmath6-graph of average degree @xmath8 . if @xmath8 is large , there is a collection of sets @xmath9 of subsets of @xmath3 satisfying * if @xmath10 is independent , there is some @xmath11 with @xmath12 , * @xmath44 for every @xmath14 , * @xmath45 where @xmath16 .
the proof of theorem [ thm : simplecont ] follows quite closely the proof of theorem 3.1 in @xcite , but modifications are needed to accommodate the presence of both uniform and degree measures
. however , nothing stronger than markov s inequality is needed .
the spirit of the proof is readily explained .
we need to identify a set of vertices that are not in @xmath4 ; then @xmath19 will be the remaining vertices .
we shall show that there are three small subsets @xmath46 , @xmath31 and @xmath47 of @xmath48 , such that @xmath46 , @xmath31 and @xmath47 determine such a set @xmath49 disjoint from @xmath4 .
this means that the number of different container sets @xmath19 that are so specified is at most the number of triples of small subsets @xmath50 ; this number is not large and this is where the bound on @xmath51 comes from .
how , then , can we specify @xmath46 , @xmath31 and @xmath47 in such a way as to enable us to identify a set @xmath49 of vertices not in @xmath4 ?
there are no edges with all @xmath6 vertices inside @xmath4 , but there are many edges altogether .
so there must be a number @xmath52 , @xmath53 , such that there are significantly fewer edges with @xmath54 vertices in @xmath4 than there are edges with @xmath52 vertices in @xmath4 .
we might then expect to find a substantial set @xmath55 of vertices each lying in many of the latter kind of edges .
so we pick small subsets @xmath56 and @xmath57 at random , and look at the set @xmath58 where @xmath59 , etc .
notice that @xmath60 is determined by @xmath46 and @xmath31 .
if we write @xmath61 then clearly @xmath62 is a container for @xmath4 that is specified by @xmath50 .
now @xmath46 and @xmath31 are small by definition , and we expect @xmath47 also to be small , because there are few edges with @xmath54 vertices in @xmath4 . on the other hand , vertices of @xmath63 have a good chance of lying inside @xmath60 , so we expect @xmath60 to contain much of @xmath63 and so have substantial measure , meaning that @xmath64 is bounded away from one .
this is the heart of the proof .
let @xmath65 be the vertex set of @xmath1 of size @xmath66 and @xmath67 the edge set . for sets @xmath68 and @xmath69 , let @xmath60 be as defined above .
given subsets @xmath70 , let @xmath71 note that @xmath72 by definition .
we will show that for every independent set @xmath4 , there are small subsets @xmath73 such that @xmath74 .
specifically , let @xmath75 note that @xmath76 is small if @xmath8 is large ( depending on @xmath6 ) .
we now define the collection @xmath2 by @xmath77 for @xmath8 sufficiently large , where @xmath16 .
this collection @xmath9 will satisfy the conditions of the lemma .
fix an independent set @xmath4 .
for a subset @xmath78 with @xmath79 , and for @xmath80 , we define the set of edges @xmath81 let @xmath82 be the statement @xmath83 statement @xmath84 is true by ( [ eqn : mubig ] ) , since @xmath85)$ ] .
statement @xmath86 is false , because @xmath4 is independent and so @xmath87 .
there must therefore exist @xmath88 such that @xmath82 is true and @xmath89 is false . fix a set @xmath90 witnessing the falsity of @xmath89 ; thus @xmath79 , @xmath91 and @xmath92 .
for @xmath93 , let @xmath94 let @xmath95 . note that @xmath96 .
consider an edge @xmath97 with @xmath98 . then @xmath99 and @xmath100 .
since @xmath101 we can pick @xmath102 with @xmath103 .
now @xmath104 so , by definition of @xmath105 , we have @xmath106 .
moreover , since @xmath103 , the definition of @xmath63 and the fact that @xmath107 imply @xmath108 .
therefore , the total number of edges in @xmath109 but not in @xmath110 is less than @xmath111 . by the choice of @xmath90 as witness set , we know that @xmath92 and so @xmath112 . since @xmath82 is true , this means @xmath113 . but @xmath91 and therefore @xmath114 .
let @xmath115 , so @xmath116 .
since @xmath117 , we observe that @xmath118 let @xmath119 and @xmath120 be random sets where each vertex ( of @xmath4 and @xmath121 respectively ) is included independently with probability @xmath122 . by markov s inequality , the inequalities @xmath123 and @xmath124 each hold with probability at least @xmath125 .
let @xmath126 .
then clearly , @xmath127 provided @xmath128 .
so to complete the proof , it is enough to show that the inequalities @xmath129 and @xmath128 each hold with probability at least @xmath125 , because then , with positive probability , all four inequalities @xmath130 and @xmath131 will hold . a vertex @xmath132 will be included in @xmath60 ( i.e. , in @xmath47 ) if it lies in an edge @xmath133 with @xmath134 , @xmath135 , @xmath136 .
therefore @xmath137 and @xmath138 , which means @xmath139 .
we know @xmath92 .
for an edge @xmath140 with @xmath138 , there are @xmath54 partitions of @xmath133 of the form @xmath141 with @xmath142 , @xmath143 and @xmath144 . for each such partition , the probability that both @xmath135 and @xmath145 is @xmath146 .
so the expected size of @xmath47 is at most @xmath147 applying markov s inequality again implies that @xmath148 with probability at least @xmath125 .
recall that @xmath149 by definition of @xmath63 , and so in particular @xmath150 .
let @xmath151 .
then @xmath152 , and so @xmath153 .
let @xmath154 .
then @xmath155 ( since @xmath156 is small ) .
each @xmath106 has a partition @xmath157 with @xmath158 and @xmath159 , where @xmath160 because @xmath103 .
the probability that @xmath161 and @xmath162 is @xmath146 and , because @xmath1 is simple , these events over all @xmath106 are independent .
hence the probability that @xmath163 , that is , @xmath164 , is at most @xmath165 now @xmath166 where @xmath167 is the indicator of the event @xmath163 . taking expectations , @xmath168 , since @xmath169 .
markov s inequality implies that , with probability at least @xmath125 , @xmath170 holds , and hence @xmath171 .
this completes the proof .
we remark that the proof shows the theorem to be true for a smaller value of @xmath172 , namely @xmath173 for some function @xmath174 of @xmath6 , but the results of @xcite are better still , with @xmath175 , so we keep the present value for simplicity .
one might wonder why the bound here on @xmath176 is worse than the bound in @xcite .
it is not because of the random choice of @xmath46 and @xmath31 , because in the context of the present algorithm random choice is quite efficient , and a deterministic choice is unlikely to be much better .
the reason that the present method is relatively inefficient is that it uses edges with exactly @xmath52 vertices in @xmath4 for one value of @xmath52 only , and ignores all other edges .
the methods of @xcite and @xcite , which are unrelated to the present method , are not lengthy to describe but are nonetheless crafted carefully to use all edges and to be as efficient as possible .
applying theorem [ thm : simplecont ] repeatedly , as described earlier , we obtain the main theorem . as we remarked earlier ,
let us apply theorem [ thm : simplecont ] to @xmath1 itself , and then again to each container so obtained , then to each of the new containers , and so on for each container with at least @xmath177 edges , until we obtain a collection @xmath2 of containers @xmath19 with @xmath17 ) < \delta e(g)$ ] . since , by ( [ eqn : musmall ] ) , each application of theorem [ thm : simplecont ] decreases the fraction of edges by @xmath178 , @xmath2 is obtained after at most @xmath179 levels of iteration , and so @xmath180 , where @xmath172 is the maximum over all applications of theorem [ thm : simplecont ] .
if @xmath17 ) \ge
\delta e(g)$ ] then the average degree of @xmath20 $ ] is at least @xmath181 , and the result follows , provided , for the sake of a clean statement , the reader will indulge us by taking the value @xmath182 rather than the weaker bound explicit in theorem [ thm : simplecont ] .
as remarked earlier , for regular hypergraphs , the condition @xmath17 ) < \delta e(g)$ ] implies @xmath183 . plugging this value into theorem 2.1 of @xcite
immediately improves by a factor of two the bound on the list colouring number in theorem 1.1 of @xcite ; nevertheless the bound obtained remains a factor of two worse than the bound in theorem 1.3 of @xcite , which is probably best possible .
we now give an application of theorem [ thm : sparsecont ] in a situation where the hypergraph of interest is not simple . in what follows , @xmath0 is a fixed @xmath184-graph .
we call another @xmath184-graph _ @xmath0-free _ if it has no subgraph isomorphic to @xmath0
. the maximum size of an @xmath0-free @xmath184-graph on @xmath185 vertices is denoted by @xmath186 , and @xmath187 is the limiting maximum density of @xmath0-free @xmath184-graphs .
the meaning of this theorem is that every @xmath0-free @xmath184-graph is a subgraph of one of just a few @xmath184-graphs that are nearly @xmath0-free .
the strength of the result can be measured by the bound on @xmath196 is some positive constant is worthwhile .
it is not possible for @xmath197 to exceed the value @xmath198 , and in fact a best possible bound was obtained in ( * ? ? ?
* theorem 1.3 ) , but the method here is simpler .
any value of @xmath199 , such as that given by theorem [ thm : ffree_cover ] , immediately gives the following corollary , because each graph @xmath19 in the statement of the theorem has at most @xmath200 subgraphs .
this corollary is the same as ( * ? ? ?
* corollary 1.4 ) . in the case
@xmath202 , this corollary was proved for complete @xmath0 by erds , kleitman and rothschild @xcite and for general @xmath0 by erds , frankl and rdl @xcite .
nagle , rdl and schacht @xcite proved it for general @xmath184 using hypergraph regularity methods .
the present paper offers the simplest known proof .
theorem [ thm : sparsecont ] can , in a similar way , be used to give a simple way to count the number of @xmath184-graphs which have no _ induced _ copy of @xmath0 , and more generally to evaluate the probability that a random @xmath184-uniform hypergraph @xmath203 contains no induced copy of @xmath0 . for @xmath202 , the value when @xmath204 was determined by prmel and steger @xcite and for general @xmath122 by bollobs and thomason @xcite
( see also marchant and thomason @xcite ) . for general @xmath184 the value for @xmath204
was given by dotson and nagle @xcite , again using hypergraph regularity techniques .
we do nt give details of the result , which is identical to ( * ? ? ?
* theorem 2.5 ) .
we merely point out that it can be derived from a container theorem , as demonstrated in @xcite , and that the container theorem presented here can be used instead , via an argument very similar to the one used to prove theorem [ thm : ffree_cover ] .
[ thm : ffree_sparse ] let @xmath0 be an @xmath184-graph and let @xmath205 . for some @xmath206 , for @xmath185
sufficiently large and for @xmath207 , where @xmath195 , the following event holds with probability greater than @xmath208 : @xmath209 a stronger version of this corollary , with @xmath210 , was conjectured by kohayakawa , uczak and rdl @xcite ; it was proved in the case of strictly balanced @xmath0 by conlon and gowers @xcite and in full generality by schacht @xcite .
the strong version follows easily from ( * ? ? ?
* theorem 1.3 ) , as shown in @xcite , and the same argument gives corollary [ thm : ffree_sparse ] from theorem [ thm : ffree_cover ] , so we do not give details here .
we remark that the point of the corollary is how small the value of @xmath122 can be made : szemerdi s regularity lemma allows @xmath211 .
we note that kohayakawa , rdl and schacht @xcite and szab and vu @xcite both proved the corollary for complete 2-graphs with @xmath212 ( slightly better in the case of @xcite ) , but again we believe the present proof is the shortest for some @xmath199 .
it yields in a similar fashion weak versions of the other so - called kr conjectures .
the proof of theorem [ thm : ffree_cover ] consists of finding a set of containers for the independent sets in the hypergraph @xmath213 , which is defined as follows .
the @xmath214 vertices of @xmath1 are the @xmath184-sets in @xmath190 $ ] , that is , @xmath215^{(\ell)}$ ] .
the edges of @xmath1 are the subsets of size @xmath216 of @xmath3 that form an @xmath184-graph isomorphic to @xmath0 . given a subset @xmath217
, we can regard @xmath31 as the edges of an @xmath184-graph with vertex set @xmath190 $ ] .
the subset @xmath31 is independent in @xmath1 if and only if @xmath31 , regarded as an @xmath184-graph , is @xmath0-free .
a set of containers @xmath2 for the independent sets in @xmath1 is thus a set of @xmath184-graphs on vertex set @xmath185 such that every @xmath0-free graph is a subset of one of these container graphs .
thus theorem [ thm : ffree_cover ] is a statement about the existence of a collection @xmath2 of containers for @xmath218 having certain properties .
the stronger ( * ? ? ?
* theorem 1.3 ) was obtained by applying a container result directly to @xmath218 . here
, we use the simpler theorem [ thm : sparsecont ] to give a set of containers with slightly weaker properties
. we can not apply theorem [ thm : sparsecont ] directly to @xmath218 because this hypergraph is not simple .
we therefore apply it instead to a subgraph @xmath219 of @xmath218 .
each independent set of @xmath218 is independent in @xmath219 , so containers for @xmath219 will also be containers for @xmath218 . to show that these containers have the properties claimed in theorem [ thm : ffree_cover ] , we need the following lemma . [
lem : sparse ] let @xmath220 and @xmath221 .
then , if @xmath185 is large enough , there exists a simple sub - hypergraph @xmath222 of @xmath213 , such that @xmath223 and @xmath224 has average degree at least @xmath225 .
moreover , for all @xmath217 , if @xmath34 ) \ge
\eta e(g)$ ] , then @xmath226 ) \ge
\eta e(g_{\rm simple})/2 $ ] .
[ prop : supersaturation ] let @xmath0 be an @xmath184-graph and let @xmath227 .
there exists @xmath228 and @xmath229 such that if @xmath19 is an @xmath184-graph on @xmath230 vertices containing at most @xmath231 copies of @xmath0 then @xmath232 .
let @xmath188 be as given in the conditions of the theorem .
then let @xmath220 be given by proposition [ prop : supersaturation ] .
we may of course assume that @xmath233 .
choose @xmath234 so that @xmath235 . then apply lemma [ lem : sparse ] to obtain @xmath224 . apply theorem [ thm : sparsecont ] to @xmath236 with @xmath237 , with @xmath214 and @xmath238 , so @xmath8 is large if @xmath185 is large , to obtain a collection @xmath239 for the independent sets in @xmath224 .
as remarked before , every @xmath0-free @xmath184-graph @xmath4 on vertex set @xmath190 $ ] is an independent set in @xmath224 and is therefore contained in some subset @xmath240 , which itself can be regarded as an @xmath184-graph on vertex set @xmath190 $ ] .
we have @xmath241 where @xmath242 . since @xmath235 we have @xmath243 , as claimed . all that remains , then , is to verify the second assertion of the theorem . in the assertion , the number of copies of @xmath0 in @xmath19 is the same as @xmath17)$ ] . by theorem [ thm : sparsecont ] , @xmath244 )
< \delta e(g_{\rm simple})$ ] . since @xmath237 , lemma [ lem : sparse ] shows @xmath17 )
< \eta e(g ) \le \epsilon e(g)$ ] .
now @xmath245 is the number of copies of @xmath0 with vertices in @xmath190 $ ] and so @xmath246 .
so proposition [ prop : supersaturation ] implies @xmath232 , completing the proof .
we form @xmath224 by randomly choosing edges of @xmath1 and then deleting a few so that the result is a simple hypergraph .
observe that @xmath247 , so that @xmath248 where @xmath249 .
we may assume that @xmath0 has more than one edge , and so @xmath250 .
call a pair @xmath251 of edges of @xmath1 with @xmath252 , an _
overlapping pair_. notice that the number of overlapping pairs is the number of copies @xmath253 of @xmath0 with vertices in @xmath190 $ ] that have at least two @xmath184-edges in common : @xmath0 and @xmath254 must share at least @xmath255 vertices and so the number of overlapping pairs is @xmath256 .
pick a number @xmath257 with @xmath258 .
let @xmath259 be a subgraph of @xmath1 formed by picking edges independently and at random with probability @xmath260 .
let @xmath261 be the number of edges of @xmath259 .
we make use of standard bounds on the tail of the binomial distribution , to wit , if @xmath262 then @xmath263 , and
the same bound holds for @xmath264 ( see for example ( * ? ? ?
* corollary 2.3 ) ) .
so if @xmath90 is the event @xmath265 then @xmath90 holds with high probability , certainly more than @xmath266 .
observe that if @xmath90 holds then @xmath259 has @xmath267 edges . for each @xmath217 ,
let @xmath272 be the event that both @xmath34 ) \ge
\eta e(g)$ ] and @xmath273 ) \le 3pe(g[s])/4 $ ] hold .
then the probability that @xmath272 occurs is at most @xmath274)/40)= \exp(-p\theta(n^h))=\exp(-\theta(n^{\ell+\rho'}))$ ] .
let @xmath19 be the event that @xmath272 does not hold for any @xmath217 .
there are @xmath275 subsets @xmath31 , so the probability that @xmath19 fails to hold is at most @xmath276 .
there is therefore a positive probability that @xmath90 , @xmath270 and @xmath19 all hold .
let @xmath259 be a graph for which they all do hold , remove an edge from each overlapping pair , and call the result @xmath224 .
this graph has no overlapping pairs and so is simple .
the number of edges is @xmath277 .
since @xmath90 and @xmath270 hold , we have @xmath278 and @xmath279 , so @xmath280 and @xmath281 . the graph has fewer than @xmath282 vertices and so its average degree exceeds @xmath283 . finally , let @xmath217 be such that @xmath34 ) \ge
\eta e(g)$ ] .
the event @xmath19 holds , and so @xmath272 does not ; thus @xmath273 ) \ge
3pe(g[s])/4\ge 3p\eta e(g)/4 \ge ( 3/4)\eta(4/5 ) e$ ] , the last inequality holding because @xmath90 holds . therefore @xmath284 )
\ge 3\eta e/5-f$ ] .
but @xmath280 so @xmath226 ) \ge
3\eta e/5-f\ge \eta(e - f)/2=\eta e(g_{\rm simple})/2 $ ] , which completes the proof . | we give an easy method for constructing containers for simple hypergraphs .
some applications are given ; in particular , a very transparent calculation is offered for the number of @xmath0-free hypergraphs , where @xmath0 is some fixed uniform hypergraph . |
wavelets and frames have become a widely studied tool in mathematics and applied science .
one of the obvious questions is the construction of wavelets with given properties .
the classical wavelet system on the line is given by a function @xmath1 , such that the dyadic dilates and translates of @xmath2 form an orthonormal basis for @xmath3 .
thus , @xmath4 with @xmath5 , is an orthonormal basis for @xmath3 .
there are several obvious generalizations : one can replace @xmath6 by any integer @xmath7 ; one can allow several wavelet functions @xmath8 ; and one can consider an orthonormal basis for a closed subspace of @xmath3 .
there have also been several publications of wavelets in higher dimensions , cf @xcite to name few .
one of the differences in higher dimensions is that we now have many more choices in the sets of dilations and translations .
so , to fix the notation , let @xmath9 and @xmath10 be countable sets .
a @xmath11-_wavelet _ is a square integrable function @xmath2 with the property that the set of functions @xmath12 forms an orthonormal basis for @xmath13 .
the set @xmath14 is called the _
dilation set _ and
the set @xmath15 is called the _
translation set_. if we replace @xmath13 in the above definition by @xmath16 for some measurable subset @xmath17 , @xmath18 , we get a @xmath11-__s__ubspace wavelet . here
@xmath19 stands for the fourier transform @xmath20 we will often write @xmath21 for the fourier transform of @xmath22 . the most natural starting point is to consider groups of dilations and full rank lattices as translation sets
. the simplest examples would then be groups generated by one element @xmath23 , see @xcite and the reference therein . in @xcite
more general sets of dilations were considered , and in general those dilations do not form a group .
even more general constructions can be found in @xcite . in this article
, we consider a special class of wavelets corresponding to _
wavelet sets_. those are wavelet functions @xmath2 such that @xmath24 , for some measurable subset @xmath25 of @xmath26 . the wavelet property is then closely related to geometric properties of the set @xmath25 , in particular spectral and tiling properties of @xmath25 .
the study of wavelet sets then becomes an interplay between group theory , geometry , operator theory and analysis , cf @xcite .
our main results are existence theorems for such wavelets for some special dilation sets @xmath14 which are _ not necessarily _ groups , see theorem [ th-2.1 ] and theorem [ th-2.2 ] .
the article is organized as follows . in section [
s : background ] we recall some well known facts on wavelet sets and their spectral and tiling properties . in section 2
we give a construction of a special type of subspace wavelet set and we continue with two examples of such sets in section 3 . in section 4
we discuss the general framework that has motivated the construction in this article .
those are the prehomogeneous vector spaces .
in this section we recall some basic facts about wavelet sets . in particular we discuss their relation to _ spectral sets _ and to _ tilings_. a standard reference is the work by y. wang @xcite .
see also the discussion in @xcite and the reference therein .
set @xmath27 .
if @xmath25 is a lebesgue measurable subset of @xmath26 , then @xmath28 denotes the measure of @xmath25 with respect to the standard lebesgue measure on @xmath26 .
a set @xmath29 with @xmath30 is a spectral set if there exists a set @xmath10 such that @xmath31 is an orthogonal basis for @xmath32 .
the set @xmath15 is called a spectrum of @xmath25 , and @xmath33 is said to be a spectral pair .
a measurable tiling of a measure space @xmath34 is a countable collection of subsets @xmath35 of @xmath36 , such that @xmath37 for @xmath38 , and @xmath39 let @xmath40 be a measurable set with @xmath41 .
let @xmath42 and @xmath43 . 1 .
we call @xmath44 a multiplicative tiling set of @xmath36 if there exists a measurable set @xmath45 , @xmath46 , such that @xmath47 is a measurable tiling of @xmath36 .
the set @xmath48 is called a multiplicative @xmath49-tile for @xmath36 or simply a _ multiplicative tile _ ; 2 .
we call @xmath15 an additive tiling set of @xmath50 if there exists a measurable set @xmath51 , @xmath52 , such that @xmath53 is a measurable tiling of @xmath50 .
the set @xmath48 is called an _ additive @xmath15-tile _ or simply an _ additive tile _ ; 3 . a set @xmath48
is called a @xmath11-_tile _ for @xmath36 if @xmath25 is a @xmath44-multiplicative tile for @xmath36 and a @xmath15-additive tiling set for @xmath26 .
if @xmath54 then we will simply speak of multiplicative tiling and @xmath55-tile , without the reference to @xmath36 .
[ re-1 ] note , that we do not assume that @xmath56 or even @xmath57 , such that @xmath58 . that will always be the case if @xmath59 .
it should be noted , that we can always , without loss of generality , assume that @xmath60 and @xmath61 . to see that , fix a @xmath62 and replace @xmath49 by @xmath63 and @xmath25 by @xmath64 ( up to a set of measure zero ) .
note , that if @xmath49 is a group , then @xmath61 , and ( [ eq-1 ] ) follows from the tiling property ( 1 ) . the spectral property of a set is closely related to the tiling property , in particular if the spectrum is a lattice .
this was first noticed by fuglede in @xcite . for a non - empty subset @xmath65 , set @xmath66 if @xmath67 is a lattice , then so is @xmath68 . in that case @xmath68
is called _ the dual lattice _ of @xmath67 .
assume that @xmath15 is a lattice .
then @xmath48 is a spectral set with spectrum @xmath15 if and only if @xmath69 is a measurable tiling of @xmath50 .
this result and several examples led fuglede to conjecture , cf .
@xcite : a measurable set @xmath25 , with positive and finite measure , is a spectral set if and only if it is an additive tile .
several people worked on this conjecture and derived important results and validated the conjecture for some special cases , see @xcite and the references therein .
however , in 2003 , t. tao @xcite showed that the conjecture is false in dimension @xmath70 and higher if the lattice hypothesis is dropped .
the other direction was disproved by m. n. kolountzakis and m. matolcsi in @xcite .
but even now , after the spectral - tiling conjecture has been proven to fail in higher dimensions , it is still interesting and important to understand better the connection between spectral properties and tiling in particular , because of the connection to wavelet sets .
let @xmath71 be an integer .
then there exists a compact set @xmath72 of positive measure such that @xmath73 admits an orthonormal basis of exponentials @xmath31 for some @xmath74 , but such that @xmath48 can not tile @xmath26 by translation .
in particular , fuglede s conjecture is false in @xmath50 for @xmath71 .
we now discuss briefly some results by dai , larson , and speegle @xcite . for that assume
that @xmath75 and @xmath15 are countable groups acting on a measure space @xmath34 by measurable automorphisms .
we think of @xmath75 as `` dilation '' group and @xmath15 as translations .
we write correspondingly the action of @xmath75 by @xmath76 and the action of @xmath15 by @xmath77 .
two sets @xmath78 and @xmath79 are said to be @xmath75-_dilation congruent _ , @xmath80 , if there exist measurable partitions @xmath81 and @xmath82 of @xmath78 and @xmath79 , respectively , such that @xmath83 for some @xmath84 .
similarly , two sets @xmath78 and @xmath79 are said to be @xmath15-_translation congruent _ , @xmath85 , if there exist measurable partitions @xmath81 and @xmath82 of @xmath78 and @xmath79 , respectively , such that @xmath86 for some @xmath87 . in 1996 ,
the above three authors showed in @xcite , that wavelet sets exist for groups of dilations .
they first introduce the notion of _ abstract dilation - translation pair_. let @xmath36 be a metric space and @xmath49 and @xmath67 discrete groups of automorphisms of @xmath36 .
a pair @xmath88 is called an abstract dilation - translation pair if the following holds : 1 .
for each bounded set @xmath78 and each open set @xmath79 there exist @xmath89 and @xmath90 such that @xmath91 ; 2 .
there is a fixed point @xmath92 for @xmath44 such that for any neighborhood @xmath7 of @xmath92 and for any bounded set @xmath78 , there is an element @xmath89 such that @xmath93 . in @xcite
the following is proved : [ th - dls1 ] let @xmath36 be a metric space and @xmath88 an abstract dilation - translation pair with @xmath94 a @xmath44 fixed point as above . if @xmath78 and @xmath79 are bounded measurable sets in @xmath36 such that @xmath78 contains a neighborhood of @xmath94 and @xmath79 has nonempty interior and is bounded away from @xmath94 , then there exists a measurable set @xmath95 , @xmath96 which is @xmath44-dilation congruent to @xmath79 and @xmath15-translation congruent to @xmath78 .
we now apply this abstract result to @xmath26 .
if @xmath97 , @xmath98 , and @xmath99 is measurable , set @xmath100 note that the fourier transform of @xmath101 is given by @xmath102 let @xmath40 be measurable , @xmath103 , and @xmath104 .
let @xmath65 be discrete .
then a measurable set @xmath61 is called an @xmath36-subspace @xmath105-wavelet set if the set of functions @xmath106 is an orthogonal basis for @xmath107 , where @xmath108 .
[ re-2 ] note again that we do not assume that @xmath109 , nor that @xmath110 up to set of measure zero , but this follows if @xmath60 . as in remark [ re-1 ] one can always assume this by replacing @xmath49 by @xmath111 and @xmath25 by @xmath112 for a fixed @xmath113 .
we get from theorem [ th - dls1 ] : [ th - dls ] let @xmath114 be an expansive matrix , and let @xmath17 be a measurable set of positive measure such that @xmath115 .
let @xmath116 and let @xmath15 be a full rank lattice .
then there exists an @xmath36-subspace @xmath105-wavelet set .
there are several generalizations of this theorem .
we refer to @xcite for discussion and references .
we will only mention two important results here : [ thm8 ] let @xmath36 be a measurable subset of @xmath50 , with positive measure satisfying @xmath117 , for some expansive matrix @xmath114 , and let @xmath15 be a full rank lattice .
then there exists a set @xmath118 such that @xmath119 is a measurable tiling of @xmath120 and @xmath121 is a measurable tiling of @xmath36 .
in particular @xmath25 is an @xmath36-subspace @xmath105-wavelet set .
[ th - wang ] let @xmath9 and @xmath10 .
let @xmath51 be measurable , with positive and finite measure .
if @xmath25 is a measurable @xmath122-tile and @xmath33 is a spectral pair , then @xmath48 is a @xmath11-wavelet set .
conversely , if @xmath48 is a @xmath11-wavelet set and @xmath123 , then @xmath25 is a measurable @xmath122-tile and @xmath33 is a spectral pair .
let us sketch some of the ideas of the proof to underline the connection between spectral properties , tiling , and wavelet sets .
let @xmath124 . as the fourier transform is an unitary isomorphism
, it follows , that the set @xmath125 is an orthogonal basis for @xmath126 if and only if the set @xmath127 is an orthogonal basis for @xmath128 . here , as before
, we have set @xmath129 a simple calculation shows that @xmath130 the fact , that @xmath131 is a measurable tiling of @xmath36 implies that @xmath132 the orthogonal projection onto @xmath133 is given by @xmath134 and @xmath135 the spectral property implies , that @xmath136 is an orthogonal basis for @xmath32 . as the linear map @xmath137 is a unitary isomorphism @xmath138 it follows , that the set of functions @xmath139 is an orthogonal basis for @xmath140 .
putting those two things together , implies that @xmath141 is an orthogonal basis for @xmath142 .
hence @xmath143 is an orthogonal basis for @xmath107 . we refer to @xcite for further discussion and generalizations of theorem [ th - wang ] .
in this section we discuss how to construct subspace wavelet sets using a kind of `` induction '' process , i.e. , using well known facts discussed in the previous section on smaller dilation sets acting on a smaller frequency set and then extending those to our bigger dilation set and frequency set .
we start with two simple , but important , observations .
for @xmath144 we say that the product @xmath145 is _ direct _ if @xmath146 , @xmath147 , @xmath148 , implies that @xmath149 and @xmath150 .
we state the following simple lemma , but note , that we will be using the proof more than the actual statement .
we remark , that the statements in the next few lemmas and theorems hold for more general settings , i.e. , one could replace @xmath26 by a measure space and @xmath151 by a group of automorphisms of @xmath36 .
[ th-2.1 ] let @xmath40 be measurable .
let @xmath152 be two non - empty sets , such that the product @xmath153 is direct let @xmath154 .
then there exists a @xmath44-tile @xmath25 for @xmath36 if and only if there exists a measurable set @xmath155 , such that @xmath156 is a measurable tiling of @xmath36 , and a @xmath157-tile @xmath25 for @xmath7 .
assume that @xmath25 is a @xmath49-tile for @xmath36 .
set @xmath158 assume , that there are @xmath148 such that @xmath159 .
then @xmath160 for all @xmath161 , which contradicts our assumption , that @xmath162 is a measurable tiling of @xmath36 .
hence @xmath163 is a measurable tiling of @xmath7 .
we have , up to a set of measure zero : @xmath164 assume , that there are @xmath147 such that @xmath165
. then we can find @xmath166 such that @xmath167 . as the product @xmath153 is direct , and @xmath162 is a measurable tiling of @xmath36
, it follows that @xmath146 and hence @xmath149 .
thus @xmath168 is a measurable tiling of @xmath36 . for the other direction , assume that @xmath168 is a measurable tiling of @xmath36 and that @xmath163 is a measurable tiling of @xmath7 .
then ( up to sets of measure zero ) @xmath169 assume that @xmath170
. then there are unique @xmath147 , and @xmath166 such that @xmath171 and @xmath172 .
hence @xmath173 which implies that @xmath149 , as @xmath168 is a measurable tiling for @xmath36 .
but then @xmath159 , which implies that @xmath150 .
hence @xmath174 .
this shows , that @xmath162 is a measurable tiling of @xmath36 .
we would like to remark at this point , that we do not assume that @xmath61 , nor that @xmath175 .
but this will in fact be the case in most applications because @xmath49 will contain the identity matrix .
recall also from remark [ re-1 ] and remark [ re-2 ] that we can always assume that @xmath59 and @xmath61 up to set of measure zero .
the same remarks hold for the following theorems .
[ th-2.2 ] let @xmath176 be two non - empty subsets such that the product @xmath177 is direct .
let @xmath178 .
assume that @xmath17 with @xmath103 , is measurable .
let @xmath179 be discrete .
then there exists a @xmath180-wavelet set @xmath109 for @xmath36 if and only if there exists a @xmath181-tiling set @xmath182 and a @xmath183-wavelet set @xmath184 for @xmath7 .
set @xmath185 and @xmath186 .
then the conditions in lemma [ th-2.1 ] are satisfied .
assume that @xmath109 is a @xmath187-wavelet set for @xmath36 .
as above we set @xmath188 then , as above , we see that @xmath168 is a measurable tiling of @xmath36 .
as @xmath25 is a spectral set , it follows from theorem [ th - wang ] that @xmath25 is a @xmath189-wavelet set for @xmath7 .
assume now that @xmath7 is a @xmath181-tiling for @xmath36 , and that @xmath184 is a @xmath183-wavelet set for @xmath7 .
then , in particular @xmath184 is a @xmath190-tile for @xmath7 . as @xmath168 is a measurable tiling of @xmath36
, it follows that @xmath191 is a measurable tiling of @xmath36 . as @xmath184 is a spectral set it follows from theorem [ th - wang ]
that @xmath184 is a @xmath180-wavelet set for @xmath36 .
recall that if @xmath192 , and @xmath193 is a group that acts on @xmath49 from the right , then there exists a subset @xmath194 , such that @xmath195 and the product is direct .
note , that we do not assume that @xmath196 .
let @xmath197 and @xmath40 measurable with @xmath103 .
let @xmath65 be discrete .
assume that @xmath198 is a group that acts on @xmath49 from the right .
let @xmath194 be such that @xmath195 as a direct product . then there exists a @xmath105-wavelet set @xmath25 for @xmath36 if only only if there exists a @xmath199-tiling set @xmath7 for @xmath36 and a @xmath200-wavelet set @xmath184 for @xmath7 .
this follows directly from theorem [ th-2.2 ] with @xmath201 and @xmath202 .
the question is then , how to obtain a wavelet set for the starting subset @xmath7 .
the following gives one way to do that .
[ th - exwave ] let @xmath203 be a measurable set , @xmath103 .
let @xmath204 be an expansive matrix and @xmath205 .
assume that @xmath206 is a multiplicative tiling of @xmath36 , @xmath207 and @xmath208 .
if @xmath15 is a lattice , then there exists a measurable set @xmath209 such that @xmath210 is a measurable tiling of @xmath120 and @xmath211 is a measurable tiling of @xmath36 .
in particular , @xmath25 is a @xmath36-subspace @xmath105-wavelet set .
let @xmath212 and @xmath213 .
then @xmath190 is an abelian group that acts on @xmath206 from the right .
hence , there exists a set @xmath214 such that @xmath215 and the product is direct .
thus , the conditions in the previous theorems are satisfied .
let @xmath216 be such that @xmath217 is a measurable tiling of @xmath36 .
set @xmath218 .
then @xmath7 is @xmath190 invariant . by theorem [ th - dls ] there
exists a @xmath219-wavelet set @xmath25 for @xmath7 . set @xmath220 then , as before , we see that @xmath25 is a @xmath105-wavelet .
before we discuss more general situations in the next section , let us explain our results by two simple examples .
[ ex-3.1 ] for @xmath221 , let @xmath222 denote the rotation in @xmath223 by the angle @xmath94 .
let @xmath224 .
for any integer @xmath225 let @xmath226 note , that @xmath227 and that @xmath228 is a group .
let @xmath229 and @xmath230 then @xmath231 is a tiling set for the finite group @xmath232 .
as @xmath233 is expansive and @xmath234 , it follows from theorem [ th - dls ] that there exists a @xmath235-wavelet set @xmath25 for @xmath231 and hence a @xmath236-wavelet set for @xmath223 , see also lemma [ th - exwave ] .
we show here how to construct such a wavelet set .
note that we only have to construct a @xmath237-wavelet set for @xmath231 .
for that , let @xmath238\times [ 0,\tan({2\pi / m})]\,\,\ , \hbox{if } \,\ , m\neq2,4\ ] ] @xmath238 ^ 2\,\,\ , \hbox{if } \,\ , m=4\ ] ] @xmath239\times[0,1]\,\,\ , \hbox{if } \,\
, m=2\ ] ] @xmath240 it is clear that @xmath241 is an abstract dilation - translation pair with a fixed point @xmath242 . by theorem 1 in @xcite , it follows that there exists a measurable set @xmath243 such that @xmath243 and @xmath78 are @xmath15-_translation congruent _ , and @xmath243 and @xmath79 are @xmath44-_dilation congruent_. on the other hand , @xmath79 is a @xmath244-multiplicative tile and @xmath245 is a spectral pair .
it follows that @xmath243 is a @xmath244-multiplicative tile and @xmath246 is a spectral set .
thus , by theorem [ th - wang ] @xmath243 is a @xmath247 wavelet set . .
let @xmath248 , @xmath249 and @xmath229 .
note that @xmath44 is not a group anylonger .
let @xmath250 ^ 2 $ ] and @xmath251\times[0,3])\setminus[0,1]^2 $ ] .
then @xmath78 is a @xmath252-additive tile and that @xmath79 is a @xmath44-multiplicative tile .
following the same procedure , we get a set @xmath243 such that @xmath243 and @xmath78 are @xmath252-_translation congruent _ , @xmath243 and @xmath79 are @xmath44-_dilation congruent _ and
so , it follows that @xmath243 is a @xmath44-multiplicative tile and @xmath253 is a spectral set .
thus , @xmath243 is a @xmath254 wavelet set .
the wavelet set @xmath243 has the form @xmath255 the description of the set @xmath256 is as follows @xmath257 w_{2,1 } & = & a^{-2}[(0,1)\times(1,3)]\\[1ex ] \end{array}\ ] ] for @xmath258 , we have the following formulas @xmath259+(1,0)\ ] ] @xmath260.\ ] ]
the results in section [ s-2 ] give us tools to construct wavelet sets or subspace wavelet sets . in this section
we discuss some general settings for our results .
those examples were discussed in @xcite and the results therein were one of the main motivations for starting this work .
we refer to the above articles for more information and details .
the starting point is a subgroup @xmath261 such that @xmath262 has finitely many open orbits @xmath263 in @xmath26 such that @xmath264 in this case , the pair @xmath265 is called a _
prehomogeneous vector space_. furthermore , we assume that @xmath266 is compact for all @xmath267 , @xmath268 .
the reason for those assumptions is , that this implies that @xmath269 is admissible in the sense of @xcite , and we can define the continuous wavelet transform .
in fact , there exists a function @xmath270 such that @xmath271 for almost all @xmath272 .
here @xmath273 stands for a left - invariant haar measure on @xmath269 .
note , that we need the transpose of @xmath274 here , because of the action in the frequency domain , cf ( [ eq - ft ] ) .
let @xmath275 be the semi - direct product of the groups @xmath26 and @xmath269 .
this is a natural generalization of the @xmath276-group in one dimension .
define @xmath277 by @xmath278 that @xmath279 is an unitary isomorphism onto its image in @xmath280 was shown in @xcite .
it is the _ continuous wavelet transform_. the inverse of @xmath281 is given by the weak integral : @xmath282 we refer to @xcite and section 2 of @xcite for detailed discussion and proofs .
the natural question now arises to find a discrete version of the inversion formula ( [ eq - inv ] ) .
note , that we can replace @xmath283 by the subspaces @xmath284 and define a continuous wavelet transform on each of those spaces .
this is where the connection to the first part of this article comes into play .
let @xmath285 and @xmath286 acting on @xmath287 by multiplication .
then @xmath288 is just the @xmath276-group .
we have two open @xmath269-orbits @xmath289 and @xmath290 . in each case
the stabilizer of a point is trivial , i.e. just the one point subgroup @xmath291 and hence compact . in this case
the space @xmath292 is the classical hardy space of all functions in @xmath3 that can be realized as @xmath293-boundary values of holomorphic functions on the upper half - plane @xmath294 .
similarly , the space @xmath295 is the space of all @xmath293-functions on the line , which can be written as @xmath293-boundary values of holomorphic functions on the lower half - plane @xmath296 .
if we replace @xmath289 by @xmath297 , then there is only one open orbit @xmath298 .
let now @xmath299 .
then we have only one open orbit @xmath300 .
the stabilizer of a point in @xmath300 is naturally isomorphic to @xmath301 and hence compact .
a discrete version of the wavelet transform in this case would be a generalization of example [ ex-3.1 ] .
let @xmath302 be an open convex cone .
define @xmath303 and @xmath304 the cone @xmath25 is said to be homogeneous if @xmath305 acts transitively on @xmath25 and self - dual if @xmath306 .
the cone @xmath25 is symmetric if it is homogeneous and self - dual .
then there exists a solvable lie group @xmath307 , such that @xmath269 acts freely and transitively on @xmath25 .
here @xmath7 is simply connected and nilpotent , the group @xmath308 is simply connected and abelian , and @xmath308 normalizes @xmath7 .
note that @xmath309 is a one dimensional subgroup containing an expansive matrix .
furthermore @xmath310 commutes with all elements in @xmath269 .
this example is important in the study of hardy spaces on @xmath26 as well as besov - spaces associated to symmetric cones , c.f .
, @xcite 1 .
@xmath312 is compact , commutes with @xmath308 , and normalizes @xmath7 ; 2 .
@xmath308 is simply connected abelian and normalizing @xmath7 ; 3 .
the map @xmath313 is a diffeomorphism ; 4 .
there exists an expansive matrix @xmath314 which is central in the group @xmath269 ; 5 .
there exists a co - compact discrete subgroup @xmath315 .
assume now that @xmath36 is one of the open @xmath262-orbits .
for simplicity we assume that @xmath262 acts freely on @xmath36 , otherwise we will have to be a little more careful about the choice of @xmath317 .
let @xmath318 be a compact ( or precompact , measurable ) subset of @xmath7 , such that @xmath319 is a measurable tiling of @xmath7 .
let @xmath320 be a co - compact subgroup in @xmath308 .
choose compact ( or precompact measurable ) subsets @xmath321 such that @xmath322 is a measurable tiling .
finally let @xmath323 be a finite subgroup and @xmath324 a compact ( or precompact measurable ) subset such that @xmath325 is a measurable tiling of @xmath312 .
set @xmath326 and @xmath327 .
let @xmath328 , where @xmath329 .
then @xmath330 is a measurable tiling of @xmath269 ( c.f . the proof of lemma 3.2 in @xcite ) .
let @xmath114 be the expansive matrix in ( 4 ) above .
then @xmath207 and the assumptions in lemma [ th - exwave ] are satisfied .
it follows , that if @xmath67 is any full rank lattice in @xmath26 , then there exists a @xmath105-wavelet set for @xmath36 .
using the corresponding wavelet function @xmath331 allows us to write a discrete version of the inversion formula ( [ eq - inv ] ) : @xmath332 the factor @xmath333 is included to make @xmath2 into an orthonormal wavelet function .
note , that the assumption that the central element @xmath114 as above exist is not satisfied for arbitrary groups @xmath269 , even if @xmath269 has finitely many open orbits , whose union is dense .
as an example take the group @xmath334 acting on @xmath223 .
there are two open orbits , @xmath335 , and @xmath336 . here
@xmath308 is the subgroup of diagonal matrices , and @xmath7 is the subgroup of upper triangular matrices with @xmath337s on the diagonal . in this case ,
the center of @xmath269 is trivial .
note , that @xmath269 is isomorphic to the @xmath276-group , hence this is in fact an example of the @xmath276-group acting on @xmath223 .
x. dai , d. r. larson , and d. m. speegle .
wavelet sets in @xmath339 .
ii . in _ wavelets , multiwavelets , and their applications ( san diego , ca , 1997 ) _ , volume 216 of _ contemp .
_ , pages 1540 .
providence , ri , 1998 . m. n. kolountzakis and m. matolcsi .
complex hadamard matrices and the spectral set conjecture . in _ proc .
7th international conference on harmonic analysis and partial differential equations ( el escorial ) , 2004 _ , to appear .
( ca/11512 ) .
g. lafsson and d. speegle .
wavelets , wavelet sets , and linear actions on @xmath339 . in _
wavelets , frames and operator theory _ , volume 345 of _ contemp . math .
_ , pages 253281 .
soc . , providence , ri , 2004 .
g. weiss and e. n. wilson . the mathematical theory of wavelets . in _
twentieth century harmonic analysis a celebration ( il ciocco , 2000 ) _ , volume 33 of _ nato sci .
_ , pages 329366 .
kluwer acad .
publ . , dordrecht , 2001 . | most of the examples of wavelet sets are for dilation sets which are groups .
we find a necessary and sufficient condition under which subspace wavelet sets exist for dilation sets of the form @xmath0 , which are not necessarily groups .
we explain the construction by a few examples . |
the canonical partition function of the dynamical triangulation model is defined as a sum over simplicial manifolds consisting of equilateral four - simplices glued together according to triangulations @xmath0 , @xmath1 here @xmath2 is the number of triangles in the simplicial manifold consisting of @xmath3 four - simplices and the topology is chosen to be that of @xmath4 .
( for more details see @xcite . )
we recall that the system has two phases , a crumpled phase at low @xmath5 and an elongated phase at high @xmath5 .
the model is supposed to represent the quantum gravitational path integral over euclidean spacetimes weighted with the regge - einstein action , with volume @xmath6 and bare newton constant @xmath7 . as such it should be able to reproduce semiclassical einstein gravity as an effective theory .
this can presumably be investigated using coordinate invariant correlation functions .
another test is to see if scalar test particles form bound states , with appropriate binding energies under nonrelativistic conditions . in @xcite
we reported on such preliminary binding energy calculations .
more insight was needed at that time about the spacetimes described by the model . for a semiclassical interpretation in terms of an effective action it is obviously important if there is in some sense a background spacetime which resembles the classical @xmath4 . to address this question we studied @xcite the number of simplices at geodesic distance @xmath8 , @xmath9 .
we found encouraging scaling of this function in a region around the transition between the crumpled and elongated phase .
( for a different scaling analysis see @xcite . ) at the transition we observed that @xmath9 behaves approximately like a four - sphere , @xmath10 $ ] , for distances @xmath8 outside a short distance ` planckian regime ' up to intermediate distances of order @xmath11 . for larger distances fluctuations
distort the shape of @xmath9 .
this suggests that we focus our attention on this intermediate distance regime for a possible semiclassical interpretation of the data .
let @xmath12 denote the number of four - simplices winding around a triangle @xmath13 ( often called a ` hinge ' ) .
we measured correlations @xcite @xmath14 ^ 2 , \label{e1}\ ] ] where the brackets denote the average over simplicial configurations , @xmath15 with @xmath16 , @xmath12 , and similar for @xmath17 . here
@xmath18 is a ` triangle geodesic distance ' between @xmath13 and @xmath19 .
the ` disconnected part ' subtracted in ( [ e1 ] ) has the form of a correlation of the observable @xmath20 with the distance to a point @xmath19 , @xmath18 , which is itself a function of the geometry .
this at first sight unfamiliar form is needed for the correlation to vanish at larger distances .
other expressions for similar correlations are conceivable .
for example , including a volume element @xmath21 with a summation over @xmath13 may appear more natural from the regge calculus point of view .
we shall argue below that this leads to the same scaling results .
taking the average @xmath22 separately in numerator and denominator is another option , which should not make significant difference for not too large distances , because of the self - averaging effect of the summations in ( [ e2 ] ) . in regge calculus @xmath23
is related to the scalar curvature @xmath24 , and @xmath12 to a local volume element at @xmath13 ; we write therefore @xmath25 ( @xmath26 ) for @xmath27 ( @xmath28 ) .
however , generally , in a scaling region , a lattice operator is expected to reduce to a combination of various continuum operators , weighted with powers of the lattice distance according to their dimensions .
assuming that the theory can be described by a continuum metric tensor @xmath29 with corresponding curvature @xmath30 , the dominant continuum observable corresponding to both our lattice correlation functions @xmath25 and @xmath26 would be given by @xmath31@xmath32 ^ 2.\ ] ] here @xmath33 is the geodesic distance between the points @xmath13 and @xmath19 for a given metric @xmath29 , and the average over geometries is supposed to be calculable in semiclassical fashion .
of course , we do not know the effective action specifying this average
. it could be a combination of @xmath34 and higher order terms like @xmath35 .
[ f1 ] shows the two terms in eq .
( [ e1 ] ) separately for @xmath36 for @xmath37 and @xmath38 ( crumpled phase ) , 1.22 ( transition ) and 1.50 ( elongated phase ) .
the two contributions are very similar and fig .
[ f2 ] shows their difference , @xmath36 , which drops to zero before finite size effects take over .
[ f3 ] shows @xmath36 for a system of @xmath39 simplices and @xmath40 , which is at the transition between crumpled and elongated phase . for this case , with the employed definition of geodesic distance @xcite , the ` planckian regime ' mentioned in the introduction is @xmath41 , while @xmath42 .
so the approximate @xmath4 background geometry is to be found in the region @xmath43 .
we see an approximate power behavior @xmath44 for @xmath45 .
a fit in the range @xmath46 gives @xmath47 , @xmath48 .
the plot for @xmath26 looks similar and leads to a compatible power @xmath49 , with @xmath50 .
the equality of the powers supports the argument given above that @xmath25 and @xmath26 are dominated by the same operators in the scaling region .
the power law behavior of the data may be expected for a flat spacetime background , with @xmath51 or @xmath52 for one- or two - graviton exange .
the approximate @xmath4 interpretation of the spacetimes near the transition implies finite size corrections to the power law . an @xmath4 background curvature
@xmath53 would also allow a behavior @xmath54 , by dimensional reasoning .
we measured @xmath55 also in the elongated phase at @xmath56 , where the system is known to behave like a branched polymer with effective dimension @xmath57 .
[ f4 ] shows , a power law looks better in this phase and a fit in the region @xmath58 leads to @xmath59 for @xmath25 and @xmath60 for @xmath26 .
perhaps this can be interpreted in an effective conformal mode theory @xcite . in the crumpled phase
the correlation data are difficult to analyse because the distances are rather short in this phase .
we draw the optimistic conclusion from the power behavior that the correlation functions indicate the presence of massless excitations .
the behaviour of test particles in the spacetimes generated with the dynamical triangulation model is another interesting subject .
test particles have no backreaction on the geometry and a calculation of their attraction due to fluctuations is analogous to the quenched or valence approximation in lattice qcd .
let @xmath61 be a scalar field propagator on a given simplicial manifold , @xmath62 , where now @xmath63 denote the points on the dual lattice and @xmath64 is the corresponding lattice laplacian .
we measured @xmath65 ^ 2 \rangle,\end{aligned}\ ] ] where @xmath18 is the ( dual lattice ) geodesic distance between four - simplices . on a flat background
one would expect the behavior @xmath66 , @xmath67 at large distances , with @xmath68 the ( renormalized ) mass of the scalar particle and @xmath69 the mass of a two - particle state ( in the quenched approximation ) .
this is a bound state if @xmath70 , with binding energy @xmath71 .
[ f5 ] shows @xmath72 and @xmath73 for @xmath39 at the transition , @xmath40 , for several values of the bare mass , @xmath74 , 0.316 , 0.1 , 0.0316 .
the approximate @xmath4 behavior mentioned in the introduction is now found in the region @xmath75 , with the distance definition emplyed here .
we see roughly exponential behavior suggesting that the @xmath4 character of the background geometry can in a first approximation be neglected , similar to the correlation functions above .
the corresponding renormalized masses are @xmath76 , 0.57 , 0.27 , 0.14 .
the ratio @xmath77 appears to increase substantially as @xmath78 gets smaller .
if the model describes semiclassical gravity we may expect hydrogen atom like behavior in a nonrelativistic situation , with small fine structure constant @xmath79 : @xmath80 in this way we may be able to define the renormalized newton constant @xmath81 . to get a rudimentary feeling for the corrections to ( [ e3 ] ) the hamiltonian @xmath82 may be useful .
replacing @xmath83 and minimizing the energy leads to @xmath84 , which suggests that @xmath85 has to be considered ` large ' .
clearly , there is no indication of nonrelativistic @xmath87 behavior .
to judge the situation we have to make sure that @xmath88 is not too large ( assuming @xmath81 exists in the model ) , or @xmath68 too small for neglecting finite size @xmath4 corrections , or both .
it is also desirable to get a semiclassical understanding of the quenched approximation .
finally , we may have to take into consideration that the euclidean effective action at intermediate distances ( neglecting higher order terms like @xmath35 ) is unbounded .
9 s. catterall , these proceedings . b.v .
de bakker and j. smit , nucl .
b. ( proc . suppl . ) 34 ( 1994 ) 739 .
de bakker and j. smit , nucl .
b439 ( 1995 ) 239 . j. ambjrn and j. jurkiewicz , _ scaling in four dimensional quantum gravity _ , nbi - he-95 - 05 .
de bakker and j. smit , _ two - point functions in 4d dynamical triangulation _ ,
itfa-95 - 1 , to be published in nucl . phys .
b. i. antoniadis and e. mottola , phys .
d45 ( 1992 ) 2013 . | we study correlations on the euclidean spacetimes generated in monte carlo simulations of the model . in the elongated phase , curvature correlations appear to fall off like a fractional power . near the transition to the crumpled phase this power is consistent with 4 .
we also present improved data of our computations of the binding energy of test particles . |
the scattering of partial waves by a central perturbating potential is fully characterized by the scattering matrix . beyond the born approximation , this matrix , which reduces to a scalar in the absence of coupling ,
is simply expressed with the phase shift @xmath2 of the partial wave with orbital angular momentum @xmath1 . when two waves with different orbital angular momenta are coupled by a spin - orbit interaction ,
the scattering matrix can be constructed with two amplitudes @xmath3 and two phase shifts @xmath4 as shown in ref .
@xcite the generalized variable phase method@xcite is a powerful tool to obtain these four parameters : one defines two amplitude functions , @xmath5 , and two phase functions , @xmath6 , which would characterize a scattered partial wave in a potential truncated at the radius @xmath7 .
these functions satisfy a system of four coupled non - linear first - order differential equations .
the numerical solution of this system permits to obtain not only the four parameters in the asymptotic region ( @xmath7 greater than the range of the potential ) , but also the radial wave functions for each value of @xmath7 .
the generalized variable phase equations contain functions which are singular at the origin . as a consequence , application of standard numerical procedures
right from the origin fail .
in addition , the numerical problem is further complicated when the scattering potential is of the coulombic type .
therefore , one is bound to analyze the behavior of the solutions of the differential equations near the origin .
the computation of reliable solutions of the four coupled differential equations is otherwise impossible .
the analysis presented in this article originates in our work on the scattering states of coupled semiconductor valence - band holes in a point defect potential @xcite .
our work now is devoted to the development of a general framework allowing the treatment of class of problems involving coupled schrdinger equations of the form given by eqs . and .
the article is organized as follows . in section
ii , we introduce the definitions and notations we use throughout the paper .
the mathematical framework is that of the standard time - independent nonrelativistic scattering theory presented in the books of , e.g. , newton@xcite , taylor@xcite , and reed and simon@xcite . in section iii
, we see that for a singular potential only one of the regular solutions can be expanded as a power series ; the search for an expansion of the other solution near the origin using the lippmann - schwinger equation leads to a logarithmically divergent integral . in section iv , we develop a matrix - based method to treat this singularity problem .
it yields an expansion of the radial wave functions near the origin , which may contain logarithmic terms .
the presence of these terms explains the failure of the methods described in section iii .
we provide numerical examples in section v to illustrate our calculations .
we choose the widely used screened coulomb potential of the yukawa type to model the scattering of holes by an ionized impurity potential @xcite , which corresponds to a realistic situation in doped semiconductors .
explicit expressions of functions and matrices omitted in the text , and a detailed discussion of a conditioning method for the lippmann - schwinger equation are given in a series of appendices .
we consider a hamiltonian @xmath8 that characterizes the scattering process of a particle with momentum @xmath9 , by a central potential @xmath10 , where @xmath7 is the distance from the origin of the potential .
this hamiltonian also contains a spin - orbit interaction term .
we restrict our analysis to stationary scattering states of @xmath8 with positive energy @xmath11 . introducing the variable @xmath12 ,
the partial waves characterized by the total angular momentum @xmath13 , may be written as@xcite : @xmath14 where @xmath15 and @xmath16 are the orbital and spin angular quantum numbers , and @xmath17 is the projection of @xmath18 on the quantization axis .
we assume that the radial schrdinger equations satisfied by the functions @xmath19 take the form : @xmath20 u_{l-1}(x ) + { \mathcal b}_{12 } \left [ \frac{{\rm d}^2}{{\rm d}x^2 } + \frac{2l+1}{x } \frac{{\rm d}}{{\rm d}x } + \frac{(l-1)(l+1)}{x^2 } \right ] u_{l+1}(x)\\ & = & \rule{0pt}{4ex } [ v(x)-1 ] u_{l-1}(x)\end{aligned}\ ] ] and @xmath21 u_{l-1}(x ) + { \mathcal b}_{22 } \left [ \frac{{\rm d}^2}{{\rm d}x^2 } - \frac{(l+1)(l+2)}{x^2 } \right ] u_{l+1}(x)\\ & = & \rule{0pt}{4ex } [ v(x)-1 ] u_{l+1}(x)\end{aligned}\ ] ] where @xmath22 is the reduced potential @xmath23 . in the absence of the spin - orbit interaction ,
the matrix @xmath24 $ ] whose elements are @xmath25 ( @xmath26 ) , reduces to the @xmath27 identity matrix @xmath28 , and one recovers the uncoupled radial schrdinger equations associated to the two values of the orbital angular momentum : @xmath29 .
the real matrix @xmath24 $ ] is assumed to be symmetric and positive definite ; it thus may be diagonalized as follows : @xmath30 = \left[{\mathcal p}\right]\left[{\mathcal r}\right]^{-2}\left[{\mathcal p}\right]^{-1},\ ] ] where the matrix @xmath31 $ ] is a @xmath27 passage matrix , whose elements are denoted @xmath32 .
the exact knowledge of the matrices @xmath33 $ ] and @xmath31 $ ] is not required to understand the analysis developped in the present article ; we only assume that the diagonal elements of @xmath33 $ ] , @xmath34 and @xmath35 , are positive .
the schrdinger equations and can be recast as : @xmath36 \frac{{\rm d}^2}{{\rm d}x^2 } + \left(\left[{\mathcal l}\right]\left[{\mathcal b}\right]-\left[{\mathcal b}\right]\left[{\mathcal l}\right]\right ) \frac{1}{x } \frac{{\rm d}}{{\rm d}x } + \frac{1}{x^2 } \left({\mathds 1}_2-\left[{\mathcal l}\right]\right)\left[{\mathcal b}\right]\left[{\mathcal l}\right]\right\}\!\ ! \left(\!\!\begin{array}{c }
u_{l-1}(x)\\ u_{l+1}(x ) \end{array}\!\!\right ) \!=\ ! [ v(x)-1]\!\ !
u_{l-1}(x)\\ u_{l+1}(x ) \end{array}\!\!\right),\ ] ] where @xmath37=\left ( \begin{array}{c c } l & 0\\ 0 & -(l+1)\\ \end{array}\right)$ ] .
hereafter , we assume that near the origin the behavior of the reduced potential @xmath22 is well represented by a convergent series expansion of the form : @xmath38 where @xmath39 is a positive integer and @xmath40 are real coefficients .
this amounts to restrict the nature of a possible singularity to a simple pole in @xmath41 . within this framework , unscreened and screened coulomb potentials @xcite , the square potential @xcite as well as any potential with a linear behavior near the origin
may thus be treated @xcite .
the sum in the above definition of @xmath22 may , e.g. , correspond to the contribution of screening in the yukawa potential .
if we impose the equalities @xmath42 , the matrix @xmath24 $ ] is equal to the identity matrix [ see ( [ btor ] ) ] .
the differential system ( [ schmat ] ) thus reduces to : @xmath43\right)\left[{\mathcal l}\right]\right\ } \left(\!\!\begin{array}{c } u_{l-1}(x)\\ u_{l+1}(x ) \end{array}\!\!\right ) = [ v(x)-1]\!\ ! \left(\!\!\begin{array}{c }
u_{l-1}(x)\\ u_{l+1}(x ) \end{array}\!\!\right)\ ] ] given that the matrix product @xmath44\right)\left[{\mathcal l}\right]$ ] is diagonal , the radial wave functions @xmath19 are , in this case , solutions of two _ uncoupled _ schrdinger equations : @xmath45 u_l(x ) = 0,\ ] ] with @xmath46 .
the free regular solution of ( [ schmatc ] ) is a riccati - bessel function@xcite @xmath47 , which behaves like @xmath48 near the origin .
the perturbed solution can thus be expanded as : @xmath49 near @xmath41 . in ( [ soldev ] ) , the coefficients @xmath50
satisfy the recurrence relation : @xmath51 with @xmath52 arbitrary and @xmath53 .
the series ( [ soldev ] ) can also be obtained iteratively from the lippmann - schwinger equation : @xmath54 with @xmath55 where @xmath56 is a riccati - neumann function@xcite . with this method ,
the coefficients @xmath50 are obtained by replacing the functions @xmath57 , @xmath58 and @xmath59 in ( [ soliter ] ) by their expansions in ascending powers of @xmath60 and @xmath61 @xcite .
note that at each step of the iteration the integrals converge at the origin .
if the coupling in the differential system ( [ schmat ] ) is restored , it seems _ a priori _ natural to adopt the same approach as above , i.e. to search for a power series of the wave functions @xmath19 .
let @xmath62 in the above sum , there is no finite constant terms @xmath63 to ensure the regularity of the wave functions : @xmath64 . from ( [ schmat ] ) , we establish the following recurrence relation : @xmath65\right)\left[{\mathcal b}\right ] \left((n+1){\mathds 1}_2-\left[{\mathcal l}\right]\right ) \left ( \begin{array}{c } a_{n+1}\\ b_{n+1}\\ \end{array}\right ) = - \left ( \begin{array}{c } a_{n-1}\\ b_{n-1}\\ \end{array}\right ) + \sum_{p=0}^{n-1 } v_p \left ( \begin{array}{c } a_{n - p}\\ b_{n - p}\\ \end{array}\right)\ ] ]
note the similarity between ( [ cpliter ] ) and ( [ iter ] ) . the matrix product @xmath66\right)\left[{\mathcal b}\right]$ ] is invertible except for @xmath67 and @xmath68 . using ( [ fullsol0 ] ) and the matrix @xmath69 $ ] , given in appendix [ wmat ]
, we find that the regular free solutions @xmath70 of ( [ schmat ] ) take the following form : @xmath71 @xmath72 where @xmath73 and @xmath74 are two arbitrary constants , which appear in the vector @xmath75 ( see section iv.b ) .
the two values of @xmath76 for which the matrix product @xmath66\right)\left[{\mathcal b}\right]$ ] is singular thus correspond to the orders of the bessel functions appearing in the expression of the free solutions ( [ freesola ] ) and ( [ freesolb ] ) . by recurrence
we obtain : * for @xmath77 : @xmath78 and @xmath79 + * for @xmath80 : @xmath81 and @xmath82 is an arbitrary constant .
note that by taking the lowest order in @xmath60 of the right hand side of ( [ freesola ] ) we obtain : + @xmath83 * for @xmath84 : + @xmath85^{-1 } \left ( \begin{array}{cc } 2l&0\\ 0&-1\\ \end{array}\right)^{-1 } \left ( \begin{array}{c } 1\\ 0\\ \end{array}\right)\ ] ] * for @xmath86 : @xmath87 and + @xmath88/(2l+1)\ ] ] to satisfy the equation @xmath87 there are two possibilities : 1 .
@xmath89 , in which case @xmath90 and @xmath91 , and one of the constants @xmath92 or @xmath93 may be chosen arbitrarily to satisfy the following equation : + + @xmath94 .
+ + with a choice of the constants @xmath82 and @xmath92 , two regular and independent solutions that can be expanded as a power series may thus be generated .
2 . @xmath95 , which amounts to a coulombic - type behavior in @xmath41 . in this case
@xmath91 , which yields @xmath96 and @xmath90 .
the equation + + @xmath97 + + shows that only one arbitrary constant may be chosen and hence only one solution may be expanded as a power series .
this solution may be obtained by cancelling the term in @xmath98 in ( [ freesola ] ) , which amounts to impose [ cf . (
[ al ] ) ] + @xmath99 + at the lowest order this solution reads : @xmath100}{\displaystyle { \mathcal p}_{12}}~r_{\rm m}^2r_{\rm p}^2 \left ( \begin{array}{c } { \mathcal b}_{12}/2\\ -{\mathcal b}_{11}/(2l+3)\\ \end{array}\right ) x^{l+2 } + { \mathcal o}(x^{l+3})\ ] ] to understand why the solution with @xmath101 can not be expanded as a power series , we study the lippmann - schwinger integral equations satisfied by the radial wave functions @xmath19 : @xmath102 where the four functions @xmath103 , given in appendix [ gmatel ] , define the green s matrix of the unperturbed system , and @xmath104 is the heaviside step function .
more precisely , we have to determine which type of function is generated by iteration from the free solution whose behavior is given by : @xmath105 at the first iteration , after cancellation of the terms in @xmath106 , @xmath107 and @xmath108 , the integral @xmath109 generates the term @xmath110 } \left(r_{\rm p}^2 - r_{\rm m}^2\right ) x^{l+1 } = b_{l+1}x^{l+1}\ ] ] in the function @xmath111 .
note that the calculations are rather tedious because of the contributions of the second order terms of the bessel and neumann functions @xmath112 and @xmath113 . at the second iteration ,
similar calculations show that the two integrals @xmath114 generate terms proportional to @xmath115 in @xmath19 .
each of these terms contains one logarithmically divergent integral in zero : @xmath116 notice that the above constant vector is colinear to the vector @xmath117 in the solution ( [ frob ] ) , which can be expanded as a power series .
our analysis thus shows why only one solution can be expanded as a power series while logarithmic terms are expected in the other solution .
it is now instructive to mention a conditioning method developped by newton @xcite .
considering the case of a tensorial scattering potential that couples two angular momenta of values @xmath118 , newton studied the behavior of the regular solutions near the origin and presented a method to obtain a well - conditioned lippmann - schwinger integral equation to avoid the divergence problem .
a detailed description of newton s method is given in appendix [ rouge ] .
here , we can say that although newton suggested that this method may be applied in a general case @xcite , the modification of the lippmann - schwinger equation is impracticable in two particular situations : * when the difference between the values of the two coupled angular momenta is greater than 2 ; * when the coupling involves a nondiagonal green s matrix . in the present work @xmath119 and the potential that couples the two components @xmath120 is scalar , but
the green s matrix is not diagonal and diverging terms appear only at the second iteration of the lippmann - schwinger equation .
for this reason we have developed another method , presented in the next section .
the coupled radial schrdinger equations and may be written as a factor product : @xmath121 \left(\begin{array}{c } u_{l-1}(x)\\ u_{l+1}(x ) \end{array}\right),\ ] ] where @xmath122 is a differential operator : @xmath123 .
having in mind that the matrix @xmath24 $ ] can be diagonalized [ eq . ] , it is natural to introduce the two - component vector @xmath124 : @xmath125^{-2 } \left[{\mathcal p}\right]^{-1 } \left(\begin{array}{cc } \delta_{-l } & 0\\ 0 & \delta_{l+1 } \end{array}\right ) \left[{\mathcal p}\right ] \vec{u}(x),\ ] ] where the vector @xmath126 is defined by its components : @xmath127 and @xmath128 , which are related to the radial wave functions @xmath19 via : @xmath129^{-1 } \left(\begin{array}{c}u_{l-1}(x)\\u_{l+1}(x)\end{array}\right).\ ] ] it follows that the second - order differential system ( [ schroedinger ] ) can be transformed @xcite into a system of four first - order differential equations perturbed by a potential represented by a @xmath130 matrix @xmath131 $ ] : @xmath132 + \left[v(x)\right ] \right\ } \left(\begin{array}{c } \vec{u}\\ \vec{g}\\ \end{array}\right),\end{aligned}\ ] ] the matrices @xmath133 $ ] and @xmath131 $ ] are defined as follows : @xmath134 = \left ( \begin{array}{cc } \frac{\displaystyle 1}{\displaystyle x } \left[{\mathcal p}\right]^{-1 } \left[{\mathcal l}\right ] \left[{\mathcal p}\right ] & \left[{\mathcal r}\right]^2\\ \rule{0pt}{4ex } -{\mathds 1}_2 & -\frac{\displaystyle 1}{\displaystyle x}~\left[{\mathcal p}\right]^{-1 } \left[{\mathcal l}\right ] \left[{\mathcal p}\right]\\ \end{array}\right)\end{aligned}\ ] ] and @xmath135= v(x ) \left ( \begin{array}{cc } \left[o\right ] & \left[o\right]\\ \rule{0pt}{4ex } { \mathds 1}_2 & \left[o\right ] \end{array}\right),\end{aligned}\ ] ] where @xmath136 $ ] is the @xmath27 null matrix .
the free solution of , obtained for @xmath137 for all @xmath60 , is of the form@xcite @xmath138^{-1 } & \left[o\right]\\ \rule{0pt}{4ex } \left[o\right ] & { \mathds 1}_2 \end{array}\right ) \left[w(x)\right]\vec{c}_0,\ ] ] where @xmath75 is a constant four - component vector .
the @xmath130 matrix @xmath139 $ ] contains the regular and irregular free solutions .
it is given in the appendix [ wmat ] .
since the differential system is of order 1 , the lagrange method of the variation of constants @xcite can be applied : @xmath140^{-1 } & \left[o\right]\\ \rule{0pt}{4ex } \left[o\right ] & { \mathds 1}_2 \end{array}\right ) \left[w(x)\right ] \vec{c}(x),\end{aligned}\ ] ] for @xmath141 . in this case , the 4-component vector @xmath142 is not constant . in the variable phase approach
@xcite it is searched as @xcite @xmath143 where @xmath144 are the phase functions and @xmath145 are the amplitude functions .
note that for ease of notation , we omit the quantum numbers @xmath146 in the equation above .
differentiation of the vector @xmath142 with respect to @xmath60 yields a generalized form of the phase equations ( gpe ) @xcite , which constitute a differential system , which we denote @xmath147 , satisfied by the functions @xmath144 and @xmath145 . solving this system permits the computation of the wave function @xmath120 for all @xmath60 ; it also permits to find the four parameters @xmath3 and @xmath4 involved in the scattering matrix .
indeed , these parameters are given by the asymptotic values : @xmath148 and @xmath149 .
the gpe must be solved with @xmath150 as initial conditions to ensure the regularity of the radial wave functions @xmath19 at the origin .
the differential system @xmath147 contains the irregular functions @xmath56 . in some cases
, the reduced potential , @xmath22 , may diverge too when @xmath151 .
the computation of the solutions of the gpe thus requires an in - depth analysis of the behavior of the functions @xmath145 and @xmath144 near the origin , which can be deduced from those of @xmath152 and @xmath153 [ see eq . ] using a generalized frobenius method . near the origin , the matrix @xmath133 + \left[v(x)\right]$ ] in ( [ equadif ] )
can take the form of an expansion : @xmath154 + \left[v(x)\right ] = \sum_{p=0}^\infty \left[{\mathcal a}_p\right ] x^{p-1},\ ] ] where the matrices @xmath155 $ ] are given in the appendix [ amat ] .
since the matrix @xmath156 $ ] is never zero , a @xmath157 term , called singularity of the first kind , appears in the expansion of the matrix @xmath133 + \left[v(x)\right]$ ] near the origin .
the general solution of ( [ equadif ] ) near the origin can be searched for as @xcite : @xmath158^i,\ ] ] where @xmath159 and @xmath160 are integers that satisfy : @xmath161 for a @xmath130 system and @xmath162 , @xmath163 is an eigenvalue of the constant matrix @xmath156 $ ] .
the series expansion in ( [ eq66 ] ) may contain terms in @xmath164 , @xmath165 ^ 2 $ ] and @xmath165 ^ 3 $ ] . using ( [ equadif ] ) , it can be shown that the four - component constant vectors @xmath166 satisfy the following recurrence relation : @xmath167\right\}\vec{{\mathcal c}}_{i , j } + \vec{{\mathcal c}}_{i+1,j } = \sum_{n=1}^j \left[{\mathcal a}_n\right]\vec{{\mathcal c}}_{i , j - n},\ ] ] with the condition that for all @xmath160 , @xmath168 .
one method to obtain systematically the relevant powers of @xmath60 and @xmath164 is given in @xcite .
in the present work , the problem is reduced to the diagonalization and inversion of 4@xmath169 matrices .
one can check that the four eigenvalues of the matrix @xmath156 $ ] are those of the matrices @xmath37 $ ] and @xmath170 $ ] , i.e. @xmath171 . these eigenvalues correspond to the power of @xmath60 of the leading term of the expansion of the special functions @xmath172 near @xmath41 or @xmath173 @xcite , as shown in table [ tab1 ] .
.[tab1]values taken by the power @xmath163 of the leading term of the expansions of the special ( riccati - bessel or riccati - neumann ) functions @xmath174 near @xmath41 . [ cols="<,^,>",options="header " , ] solving the two coupled second order differential equations , ( [ schroedinger ] ) , satisfied by the functions @xmath176 , yields four constants of integration that correspond to the four independent solutions , of which two are regular and two are singular .
the two smallest eigenvalues correspond to two singular solutions which are disregarded by determining the vectors @xmath166 so that : @xmath177 for @xmath161 and @xmath178 , which amounts to impose that the related two constants of integration are zero .
the choice of the values of the other two constants corresponding to the regular solutions remains free .
the consequences of this choice are now discussed .
this eigenvalue corresponds to a behavior in @xmath98 , which is acceptable .
the vector @xmath166 is now associated to the term @xmath180^i$ ] .
let us consider the case @xmath181 for which the smallest power of @xmath60 in the expansion of ( [ eq66 ] ) coincides with the eigenvalue @xmath182 .
given that the coefficients @xmath166 of the irregular solutions are equal to zero , the recurrence equation , ( [ recceq ] ) , reduces to : @xmath183\right\}\vec{{\mathcal c}}_{i,0 } + \vec{{\mathcal c}}_{i+1,0 } = \vec{0},\ ] ] for @xmath184 . since @xmath185 , the above equality yields : @xmath183\right\}^4\vec{{\mathcal c}}_{0,0 } = \vec{0},\ ] ] which implies that either @xmath186 or @xmath187 is an eigenvector of the matrix @xmath156 $ ] associated to the eigenvalue @xmath182 : @xmath188^{-1 } \left(\begin{array}{c } 1\\ 0 \end{array}\right)\\ \rule{0pt}{6ex } v_0\left[{\mathcal p}\right]^{-1 } \left(\begin{array}{c } 1\\ 0\\ \end{array}\right ) \end{array}\right),\ ]
] where @xmath189 is an arbitrary constant . in both cases
, we also find that @xmath190 , which is equivalent to the absence of @xmath191^i$ ] terms and corresponds to the presence of the riccati - bessel function of order @xmath192 in @xmath193 .
for @xmath195 , the recurrence relation reads : @xmath196\right\}\vec{{\mathcal c}}_{i,1 } + \vec{{\mathcal c}}_{i+1,1 } = \left[{\mathcal a}_1\right]\vec{{\mathcal c}}_{i,0},\ ] ] from which we deduce the equation satisfied by the vector @xmath197 : @xmath198\right\}^4\vec{{\mathcal c}}_{0,1 } = \left\{(l+1){\mathds 1}_4 - \left[{\mathcal a}_0\right]\right\}^3\left[{\mathcal a}_1\right]\vec{{\mathcal c}}_{0,0}.\\\ ] ] since the matrix @xmath199 $ ] is not invertible , the above inhomogeneous matrix equation has a solution only on the condition that the vector on the right hand side of ( [ inheq ] ) is orthogonal to the kernel of the operator @xmath200\right\}^4 $ ] @xcite .
one can check that this condition is fulfilled when @xmath197 takes the form : @xmath201^{-1}\left[{\mathcal d}\right]^{-1 } \left[{\mathcal p}\right]\left[{\mathcal r}\right]^2\left[{\mathcal p}\right]^{-1 } \left(\begin{array}{c } 1\\ 0 \end{array}\right)\\ \rule{0pt}{6ex } \xi_0\left[{\mathcal p}\right]^{-1 } \left(\begin{array}{c } 1\\ 0 \end{array}\right ) \end{array}\right)\\ + b \left(\begin{array}{c } 0\\ 0\\ \rule{0pt}{6ex } \left[{\mathcal p}\right]^{-1 } \left(\begin{array}{c } 0\\ 1 \end{array}\right ) \end{array}\right),\ ] ] where the matrix @xmath202 $ ] and the constant @xmath203 are given by : @xmath204= \left(\begin{array}{cc } 1 & 0\\ 0 & 2l+2 \end{array}\right),\ ] ] and @xmath205}\right)\ ] ] the constant @xmath189 is the same that appears in the definition of @xmath206 , ( [ c00 ] ) , and @xmath207 is the other arbitrary constant that has to be introduced to obtain the general solution .
these two constants are indeed necessary to generate the two independent solutions that are regular at the origin .
once the vectors @xmath208 and @xmath197 are known , it is possible to calculate the vectors @xmath209 for @xmath210 .
note that the second vector on the right hand side of ( [ c01 ] ) is an eigenvector of the matrix @xmath156 $ ] associated to the eigenvalue @xmath211 ; as such it does not bring any contribution to the vectors @xmath209 for @xmath210 .
indeed we find : @xmath212 since @xmath213 is also an eigenvector of the matrix @xmath156 $ ] associated to @xmath211 , and @xmath214^{-1 } \left(\begin{array}{c } 0\\ 1 \end{array}\right ) \end{array}\right),\ ] ] where @xmath215 is given by : @xmath216}\ ] ] note that the last two components of the vector @xmath213 give rise to terms in @xmath217 $ ] , which are of the second order in the potential strength .
since from @xmath219 the matrices @xmath220 $ ] are invertible , the vectors @xmath166 can be determined without the need to introduce new constants . for @xmath219 , we obtain @xmath221\right\}^{-1 } \left[{\mathcal a}_1\right]\vec{{\mathcal c}}_{1,1}(a)\\ \vec{{\mathcal c}}_{2,2 } & = & \vec{{\mathcal c}}_{3,2}=\vec{0}\\ \vec{{\mathcal c}}_{0,2}(a , b ) & = & \left\{(l+2){\mathds 1}_4 - \left[{\mathcal a}_0\right]\right\}^{-1 } \left\{\left[{\mathcal a}_1\right]\vec{{\mathcal c}}_{0,1}(a , b ) + \left[{\mathcal a}_2\right]\vec{{\mathcal c}}_{0,0}(a ) - \vec{{\mathcal c}}_{1,2}(a)\right\}\end{aligned}\ ] ] it is not useful to give the explicit form of the vector @xmath222 , but we stress that a lengthy calculation yields four non - zero components .
terms in @xmath223 thus appear in the wave functions @xmath127 and @xmath128 , but they do not compromise their regularity at the origin . to check
if terms in @xmath165 ^ 2 $ ] and @xmath165 ^ 3 $ ] do occur in the series ( [ eq66 ] ) , we consider the general equations for @xmath224 : @xmath225\right\}\vec{{\mathcal c}}_{0,j } + \vec{{\mathcal c}}_{1,j } & = & \sum_{n=1}^j \left[{\mathcal a}_n\right]\vec{{\mathcal c}}_{0,j - n}\\ \left\{(l+j){\mathds 1}_4 - \left[{\mathcal a}_0\right]\right\}\vec{{\mathcal c}}_{1,j } + \vec{{\mathcal c}}_{2,j } & = & \sum_{n=1}^{j-1 } \left[{\mathcal a}_n\right]\vec{{\mathcal c}}_{1,j - n}\\ \left\{(l+j){\mathds 1}_4 - \left[{\mathcal a}_0\right]\right\}\vec{{\mathcal c}}_{2,j } + \vec{{\mathcal c}}_{3,j } & = & \sum_{n=1}^{j-2 } \left[{\mathcal a}_n\right]\vec{{\mathcal c}}_{2,j - n}\\ \left\{(l+j){\mathds 1}_4 - \left[{\mathcal a}_0\right]\right\}\vec{{\mathcal c}}_{3,j } & = & \sum_{n=1}^{j-2 } \left[{\mathcal a}_n\right]\vec{{\mathcal c}}_{3,j - n}\end{aligned}\ ] ] recalling that @xmath226 , we find by reccurence @xmath227 . the series ( [ eq66 ] )
thus does not contain terms neither in @xmath165 ^ 2 $ ] nor in @xmath165 ^ 3 $ ] .
this result allows us to express the vector @xmath228 as a function of the vectors @xmath229 , @xmath230 : @xmath231\right\}^{-1 } \left(\sum_{n=1}^{j-1 } \left[{\mathcal a}_n\right]\vec{{\mathcal c}}_{1,j - n}(a)\right),\ ] ] which we use to obtain the vector @xmath232 as : @xmath233\right\}^{-1 } \left(\sum_{n=1}^j \left[{\mathcal a}_n\right]\vec{{\mathcal c}}_{0,j - n}(a , b ) - \vec{{\mathcal c}}_{1,j}(a)\right).\ ] ] we observe that the vectors @xmath228 associated to the terms in @xmath234 , are proportional to the factor @xmath235 appearing in the expression of the vector @xmath213 , ( [ c11 ] ) .
the @xmath164 terms are thus only generated if the potentials contain a coulombic component ( @xmath95 ) and if the two differential equations satisfied by the wave functions @xmath127 and @xmath128 are coupled ( @xmath236 ) . since this factor is of second order in the potential strength , we find that it is consistent with the fact that the divergence problems appear only from the second iteration when searching for the solutions using the green s matrix @xcite .
the choice @xmath237 leads to @xmath186 and @xmath238 for all @xmath160 . as a consequence
, we see that there exists a solution whose expansion does not contain any logarithmic term . this particular expansion
is called a frobenius series @xcite and here it takes the following form : @xmath239 where the vectors @xmath232 , @xmath224 , now satisfy the following recurrence relation : @xmath240\right\}^{-1 } \left(\sum_{n=1}^{j-1 } \left[{\mathcal a}_n\right]\vec{{\mathcal c}}_{0,j - n}\right).\ ] ] the terms of the lowest order of the functions @xmath127 and @xmath128 are in @xmath241 since the two first components of the vector @xmath197 are zero [ see ( [ c01 ] ) ] ; more precisely , we obtain : @xmath242 \left[{\mathcal r}\right]^2 \left[{\mathcal p}\right]^{-1 } \left(\begin{array}{c } 0\\ \rule{0pt}{4ex } 1 \end{array}\right ) x^{l+2}\ ] ] at the lowest order of the expansion . from the definitions of the vectors @xmath206 and @xmath243 , and from the linearity of the recurrence relations satisfied by the vectors @xmath232 and @xmath228 ,
the following properties can be deduced : @xmath244 the vectors @xmath245 and @xmath246 are determined uniquely , recursively , from the series expansion of @xmath133 $ ] in @xmath41 .
moreover the vectors @xmath247 and @xmath245 satisfy the same recurrence relation .
since the first non - zero vectors satisfy the following equation : @xmath248}~\vec{{\mathcal c}}_{0,1}(0,1)\ ] ] we obtain @xmath249}~\vec{{\mathcal c}}_{0,j}(0,1),\ ] ] so that the general regular solution generated by two arbitrary constants @xmath250 can be written @xcite : @xmath251 where @xmath252 and @xmath253}~\left(\begin{array}{c } \vec{u}\\ \vec{g } \end{array}\right)_{(0,1)}\right]\ln(x).\ ] ] the vectors @xmath254 and @xmath255 are thus two independent solutions , of which the first is the logarithm - free solution . if @xmath256 , the expansion consists of a power series plus an additional term originating from the product of @xmath257 and the previous solution obtained with @xmath258 .
this @xmath164 term , studied in the theory of fuschian differential equations @xcite appears only at the second order for coulombic potentials , i.e. @xmath259 when @xmath260 .
this is consistent with the results obtained from the lippmann - schwinger equation .
the relationships between the constants @xmath250 and @xmath261 are as follows : @xmath262\left[{\mathcal r}\right]^{l+1}}{\displaystyle ( 2l+1 ) ! ! } \left(\begin{array}{c } a_0^+\\ \rule{0pt}{4ex } a_0 ^ - \end{array}\right).\ ] ] they are obtained from the comparison of the lowest order terms of the series expansion of @xmath263 in ( [ eq67 ] ) , and @xmath264 in ( [ freesola ] ) and ( [ freesolb ] ) .
one can check that for the logarithm - free solution the equality ( [ a0pm ] ) is satisfied .
from a computational viewpoint , the power series appearing in the expansion of @xmath265 must be truncated , which limits its use to a finite interval @xmath266 $ ] ; moreover @xmath267 is in general smaller than the distance where the functions @xmath145 and @xmath144 cease to vary and which is directly linked to the range of the potential @xmath22 . to obtain @xmath268 and @xmath269
, the differential system @xmath147 must be numerically solved from a value of @xmath60 , @xmath270 .
the values of @xmath271 and @xmath272 at the matching point @xmath273 are found from eqs . and .
to illustrate our work we choose the case of the spherical hamiltonian of semiconductor valence - band holes in a perturbating spherical potential @xmath10 , which reads @xcite : @xmath274 where @xmath275 is the free electron mass , and @xmath276 and @xmath277 are two material - dependent parameters .
the first term of @xmath8 in ( [ eqh ] ) is the kinetic energy .
the second term represents the spin - orbit interaction .
the operator @xmath278 is the scalar product of two irreducible tensors of rank 2 constructed from the components of the hole momentum , @xmath279 , and the spin angular momentum , @xmath280 @xcite .
the quantum numbers of the scattered partial wave are chosen as @xmath281 and @xmath282 . with this choice ,
the matrix @xmath24 $ ] is readily obtained@xcite : @xmath283= { \mathds 1}_2 - \frac{\mu}{5}\left ( \begin{array}{c c } 1 & 3\sqrt{6}\\ 3\sqrt{6 } & 4\\ \end{array}\right)\ ] ] the matrices @xmath31 $ ] and @xmath33 $ ] easily follow : @xmath284=\frac{1}{\sqrt{5}}\left ( \begin{array}{c c } -\sqrt{2 } & \sqrt{3}\\ \phantom{-}\sqrt{3 } & \sqrt{2}\\ \end{array}\right),\ ] ] and @xmath285= \left ( \begin{array}{c c } 1/\sqrt{1 + 2\mu } & 0\\ 0 & 1/\sqrt{1 - \mu}\\ \end{array}\right).\ ] ] the interaction between the ionized defect and the holes is represented by a yukawa potential : @xmath286 where @xmath287 is the charge number , @xmath288 is the bohr radius@xcite , and @xmath289 the screening length parameter .
this corresponds to the thomas - fermi approximation for the screening .
we took the following values : @xmath290 , @xmath291 , @xmath292 , and @xmath293 for the effective mass parameter of si @xcite .
the differential system @xmath147 for this case is similar to that given in reference@xcite . as discussed in sec .
iv.e , the logarithm - free solution is obtained setting @xmath237 . the amplitude , phase and radial wave functions are shown on figs . 1 , 2 , and 3 .
this can be done by taking @xmath294 and @xmath295 , for @xmath282 [ see ( [ a0pm ] ) ] . on figure [ fig1 ]
, we see that the constants @xmath74 and @xmath73 indeed correspond to the initial values of the amplitudes .
the dashed - point curves on the three figures are obtained from the expansion series , ( [ swoutlog ] ) , limited to @xmath296 terms . beyond the matching point @xmath297 ( here @xmath298 ) , the solid lines correspond to the solutions obtained by solving the differential system satisfied by @xmath145 and @xmath144 with the runge - kutta method @xcite .
some experimentation is needed to find the minimum value of @xmath299 for which the curves generated from the expansion remain unchanged in the interval @xmath300 $ ] , and match with those obtained by the runge - kutta method .
increasing the value of @xmath299 may only allow to obtain a matching over a wider range ; moreover , this presents little interest since this requires more computing time . at low energy ( @xmath301 ) , the oscillatory behavior of the solutions near the origin is enchanced @xcite and the value of @xmath297 must be reduced accordingly .
the solutions with logarithm were generated taking @xmath302 .
the curves shown on figs .
4 , 5 and 6 represent the amplitude , phase and wave functions .
we obtain a good agreement between the series expansion ( dashed lines ) and the runge - kutta method ( solid lines ) for @xmath298 and @xmath303 . in this case ,
@xmath74 and @xmath73 do not correspond to the initial values of the amplitude functions @xmath304 , see figure [ fig4 ] .
note that the range over which we performed the numerical calculations needs to be increased to obtain ralph s parameters@xcite , @xmath268 and @xmath269 .
to compute the wave functions of the scattering states in a coulombic potential , one must treat the related divergence problems with special care in the presence of coupling .
we showed that only one of the two regular solutions can be expanded as a power series contrary to cases treated in the literature@xcite .
we presented a matrix - based method that makes it possible to seek the expansion of the two solutions in a systematic way . using this method
, we showed that for one solution the expansion near the origin contains logarithmic terms .
these terms appear only at the second iteration of the born series .
this behavior , which is related to the particular form the spin - orbit interaction , explains why a well - conditioned lipmmann - schwinger integral equation is extremely difficult if not impossible to obtain . to our knowledge
the solutions of the spin - orbit - coupled schrdinger equations had not been studied before by direct numerical computation .
h. o. gratefully acknowledges partial support of the agence nationale de la recherche .
for convenience , the @xmath130 matrix @xmath69 $ ] can be written as : @xmath305=\left(\begin{array}{lr } \left[w_{11}\right ] & \left[w_{12}\right]\\ ~&~\\ \left[w_{21}\right ] & \left[w_{22}\right ] \end{array}\right),\ ] ] where the four @xmath27 blocks @xmath306 $ ] , @xmath307 $ ] , @xmath308 $ ] , and @xmath309 $ ] are given by : @xmath310 = r_{\rm p } \left(\begin{array}{lr } -{\mathcal p}_{11 } \hat{\jmath}_{l-1}(r_{\rm p}x ) & -{\mathcal p}_{11 } \hat{n}_{l-1}(r_{\rm p}x)\\ \rule{0pt}{4ex } \phantom{-}{\mathcal p}_{21}\hat{\jmath}_{l+1}(r_{\rm p}x ) & { \mathcal p}_{21 } \hat{n}_{l+1}(r_{\rm p}x)\\ \end{array}\right)\end{aligned}\ ] ] @xmath311 = r_{\rm m } \left(\begin{array}{lr } -{\mathcal p}_{12 } \hat{\jmath}_{l-1}(r_{\rm m}x ) & -{\mathcal p}_{12 } \hat{n}_{l-1}(r_{\rm m}x)\\ \rule{0pt}{4ex } \phantom{-}{\mathcal p}_{22}\hat{\jmath}_{l+1}(r_{\rm m}x ) & { \mathcal p}_{22 } \hat{n}_{l+1}(r_{\rm m}x)\\ \end{array}\right)\end{aligned}\ ] ] @xmath312 = \left(\begin{array}{cc } \hat{\jmath}_l(r_{\rm p}x ) & \hat{n}_l(r_{\rm p}x)\\ \rule{0pt}{4ex } 0 & 0\\ \end{array}\right)\end{aligned}\ ] ] @xmath313 = \left(\begin{array}{cc } 0 & 0\\ \rule{0pt}{4ex } \hat{\jmath}_l(r_{\rm m}x ) & \hat{n}_l(r_{\rm m}x)\\ \end{array}\right)\end{aligned}\ ] ]
the expressions of the four matrix elements of the green s matrix @xmath314 $ ] defined in ( [ lippmannschwinger ] ) are given below : @xmath315}~ g_{l-1,l-1}(r_{\rm p}x , r_{\rm p}x ' ) - \frac{r_{\rm m}{\mathcal p}_{12}{\mathcal p}_{21}}{\mbox{det}\left[{\mathcal p}\right]}~g_{l-1,l-1}(r_{\rm m}x , r_{\rm m}x')\ ] ] @xmath316}~ g_{l+1,l+1}(r_{\rm p}x , r_{\rm p}x ' ) + \frac{r_{\rm m}{\mathcal p}_{11}{\mathcal p}_{22}}{\mbox{det}\left[{\mathcal p}\right]}~g_{l+1,l+1}(r_{\rm m}x , r_{\rm m}x')\ ] ] @xmath317 } \left[r_{\rm p } g_{l-1,l+1}(r_{\rm p}x , r_{\rm p}x ' ) - r_{\rm m } g_{l-1,l+1}(r_{\rm m}x , r_{\rm m}x')\right]\ ] ] @xmath318 } \left[-r_{\rm p } g_{l+1,l-1}(r_{\rm p}x , r_{\rm p}x ' ) + r_{\rm m } g_{l+1,l-1}(r_{\rm m}x , r_{\rm m}x')\right],\ ] ] where @xmath319 which is a form more general than ( [ gfunc ] ) .
by putting together the two components @xmath320 and @xmath321 of the two independent regular solutions ( @xmath322 ) in the same matrix @xmath323 $ ] , the lippmann - schwinger integral equation corresponding to the tensorial scattering potential , takes the following form : @xmath324 note that the green s matrix , whose elements are the functions @xmath58 defined in ( [ gfunc ] ) , is diagonal in this case . the first term on the right hand side of ( [ eq60 ] ) is the @xmath27 matrix @xmath325 $ ] , solution of the equation at the order 0 .
the behavior of the functions @xmath326 near the origin @xmath41 is analyzed by iteration .
for the solution ( @xmath327 ) the integrals converge at all orders and the functions @xmath328 can be expanded as power series .
for the solution ( @xmath329 ) the first iteration of the born series leads to the evaluation of the following integral : @xmath330 which usually diverges because of the presence of the product @xmath331 that behaves like @xmath332 near @xmath333 .
we therefore see that problems arise when the index @xmath1 of the green s function @xmath334 is greater than that of the bessel function @xmath335 , which represents the free solution .
these difficulties are circumvented by adding a term to the lippmann - schwinger equation , ( [ eq60 ] ) , which modifies the inhomogeneity in a convenient way@xcite : @xmath336 $ ] becomes @xmath336\left({\mathds 1}_2+\left[m\right]\right)$ ] , where @xmath337 $ ] is a constant matrix appropriately chosen so that it cancels the problematic term that reads @xmath338 the appropriate matrix @xmath337 $ ] for the case considered , may be written as follows : @xmath339 = -(2l+1 ) \int_0^{x_0 } \frac{\displaystyle 1}{\displaystyle x ' } \left[v^{{\rm off}}(x')\right]~{\rm d}x',\ ] ] where @xmath340=\left(\begin{array}{cc } 0 & 0\\ v(x ) & 0\\ \end{array}\right)$ ] , and @xmath341 may be arbitrarily chosen . the integral that appears in the definition of @xmath337 $ ] in ( [ eq64 ] ) may be decomposed into two parts as : @xmath342 , to obtain the modified lippmann - schwinger equation : @xmath343 = \left[\psi^{(0)}(x)\right]\!\ !
\left({\mathds 1}_2 - ( 2l+1)\!\ !
\int_x^{x_0 } \frac{\displaystyle { \rm d}x'}{\displaystyle x'}~\left[v^{\rm off}(x')\right]\right)\\ \nonumber & & \hspace{-0.5cm}+ \int_0^{x } \left(\left[{\mathcal g}^0(x , x')\right]\left[v(x')\right]\left[\psi(x')\right ] - \frac{\displaystyle 2l+1}{\displaystyle x ' } \left[v^{\rm off}(x')\right]\right ) { \rm d}x',\end{aligned}\ ] ] this equation makes it possible to obtain equivalents to the solutions @xmath344 and @xmath345 in the vicinity of @xmath41 .
note that it is no longer necessary to assume that the integral that defines the matrix @xmath337 $ ] converges .
it also is instructive to see how the solutions @xmath344 and @xmath345 behave when the tensorial potential of ( [ eq60 ] ) is constant @xcite .
taking @xmath346 , we obtain at the lowest order , a modified solution of the form : @xmath347\left({\mathds 1}_2+\left[m\right]\right ) \!=\ !
\left ( \begin{array}{cc } \hat{\jmath}_{l-1}(x ) & 0\\ ( 2l+1)~v~\hat{\jmath}_{l+1}(x ) \ln(x ) & ~~\hat{\jmath}_{l+1}(x)\\ \end{array}\right)\ ] ] where @xmath348 is the off - diagonal term .
the second component of the solution @xmath344 contains an amplitude function @xmath349 that may diverge logarithmically .
this behavior , predicted in the theory of fuschian differential equation @xcite , does not compromise the regularity of @xmath350 $ ] in @xmath351 , since @xmath352 .
@xmath355 = \left ( \begin{array}{cc } \left[{\mathcal p}\right]^{-1 } \left[{\mathcal l}\right ] \left[{\mathcal p}\right ] & \left[o\right]\\ \rule{0pt}{4ex}v_0 { \mathds 1}_2&-\left[{\mathcal p}\right]^{-1 } \left[{\mathcal l}\right ] \left[{\mathcal p}\right]\\ \end{array}\right),\end{aligned}\ ] ] h. i. ralph , philips res
. repts . *
32 * , 160 ( 1977 ) .
f. calogero , _ variable phase approach to potential scattering _
( academic press , new york , 1967 ) .
p. bogdanski and h. ouerdane , phys .
b * 74 * , 085210 ( 2006 ) .
r. g. newton , _ scattering theory of waves and particles _
( mcgraw - hill , 1966 ) .
j. r. taylor , _ scattering theory , the quantum theory of nonrelativistic collisions _
( dover publications , inc . ,
mineola , new york , 2009 ) .
m. reed and b. simon , _ scattering theory , volume iii of methods of modern mathematical physics _
( academic press , new york , 1979 ) .
j. r. meyer and f. j. bartoli , phys .
b * 23 * , 5413 ( 1981 ) .
a. baldereschi and n. o. lipari , phys
b * 8 * , 2697 ( 1973 ) .
a. ronveaux , am . j. phys . *
37 * 135 , ( 1969 ) .
r. g. newton , j. math .
* 1 * , 319 ( 1960 ) . j. r. cox and a. perlmutter , nuovo cimento * 37 * , 76 ( 1965 )
. e. l. ince , _ ordinary differential equations _
( dover , new york , 1926 ) .
e. a. coddington and n. levinson , _ theory of ordinary differential equations _
( mcgraw - hill , new york , 1955 ) . a potential of the form @xmath360 , for which all the terms in the expansion are non - zero .
@xmath361 for @xmath362 and @xmath363 elsewhere .
m. lannoo and g. allan , solid state comm . * 33 * , 293 ( 1980 ) .
f. john , _ ordinary differential equations _
( courant institute of mathematical sciences , new - york university , 1965 ) .
j. joachain , quantum collision theory ( north - holland , amsterdam , 1983 ) .
w. greub , _ linear algebra _ ( springer verlag , 1981 ) .
p. bogdanski 2003 doctorate of science dissertation ( universit de caen , france ) . c. h. edwards and d. e. penney , _ elementary differential equations and applications _ ( prentice hall , 1989 ) .
n. o. lipari and a. baldereschi , phys .
lett . * 25 * , 1660 ( 1970 ) . | we report on our analysis of the coulomb singularity problem in the frame of the coupled channel scattering theory including spin - orbit interaction .
we assume that the coupling between the partial wave components involves orbital angular momenta such that @xmath0 . in these conditions , the two radial functions , components of a partial wave associated to two values of the angular momentum @xmath1 ,
satisfy a system of two second - order ordinary differential equations .
we examine the difficulties arising in the analysis of the behavior of the regular solutions near the origin because of this coupling .
first , we demonstrate that for a singularity of the first kind in the potential , one of the solutions is not amenable to a power series expansion .
the use of the lippmann - schwinger equations confirms this fact : a logarithmic divergence arises at the second iteration . to overcome this difficulty , we introduce two auxilliary functions which
, together with the two radial functions , satisfy a system of four first - order differential equations .
the reduction of the order of the differential system enables us to use a matrix - based approach , which generalizes the standard frobenius method .
we illustrate our analysis with numerical calculations of coupled scattering wave functions in a solid - state system . |
studies of galaxies in the local universe have demonstrated an apparent relation between the stellar mass of a spheroid and that of the super - massive black hole ( smbh ) which lies at its heart .
the relation for spheroids , with the mass of the smbh typically @xmath3% of the spheroid mass , suggests that the growth of the black hole and the surrounding galaxy are related .
this provides an important clue to understanding galaxy formation , as smbhs have been suggested as a highly efficient mechanism for regulating star formation in galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
feedback _ is an essential part of current galaxy formation models and provides a physically - motivated explanation for the form of the galaxy luminosity function @xcite .
however , observational evidence for feedback , in particular driven by activity from smbhs , is sparse , especially for the youngest and most massive galaxies for which it is expected to be most important @xcite .
modeling agn - driven feedback is more difficult than the traditional star - formation feedback recipes . in part
this is because we still lack a clear model for the early growth of the smbh , where bh bh mergers and super - eddington accretion may both play a part .
there is thus an urgent need for observations of the earliest phases of the evolution of spheroids and smbhs to guide development of theoretical models .
studies of the submillimeter galaxy ( smg ) population ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) suggest that luminous far - infrared galaxies are likely to be associated with an early - phase in the formation of massive galaxies .
the intensity of their starbursts , the resulting high metallicity , along with their large dynamical masses , high gas fractions and strong clustering @xcite are all suggestive of a close link to the formation phase of the most massive spheroids @xcite . if these are indeed young spheroids , what can we learn about their black holes ?
deep x - ray observations provide a sensitive probe of agn activity in submillimeter galaxies ( smg ) and can be used to place smbh mass constraints , even in the presence of heavy gas and dust obscuration . exploiting the unique sensitivity of the 2-ms exposure of the _ chandra _ deep field north (; * ? ? ?
* ) , @xcite study a small sample of _ spectroscopically identified _ smgs and find that most host weak agn with modest absorption - corrected x - ray luminosities .
this suggest two things : 1 ) the agns do not typically dominate the bolometric emission from smgs , and 2 ) the masses of the smbhs are likely to be @xmath4m@xmath2 ( eddington - limited smbh masses are typically @xmath5m@xmath2 ) .
the large agn fraction in the smg population ( @xmath640% for sources with @xmath7mjy ) further indicates that the smbhs are growing almost continuously throughout these vigorous , but obscured , star - formation episodes .
the stark contrast between the agn fractions in the smg and coeval uv - bright galaxy populations ( @xmath83% ; @xcite ) suggest that this joint black hole stellar growth activity is most strongly connected with the smg population @xcite . in this paper
we compare estimates for the stellar and smbh masses for a sample of 13 high - redshift proto - spheroids ( identified with far - infrared luminous galaxies detected in the submillimeter and radio wavebands ) in the region .
these all have precise spectroscopically identified redshifts and have hard x - ray emission dominated by an agn .
the stellar mass estimates are derived from optical , near - infrared and most importantly the mid - infrared observations of this region by the _
spitzer space telescope _ taken as part of the goods survey @xcite , while the smbh mass estimates are taken from @xcite based on the 2-ms _ chandra_observations .
together , these data allow us to investigate the origin of the relation for young spheroids in the early universe . throughout
, we adopt a cosmology with h@xmath9kms@xmath10mpc@xmath10 , @xmath11 , and @xmath12 .
unless otherwise stated , masses and luminosities are given in solar units .
for local galaxies , spatially resolved kinematics ( e.g. * ? ? ? * ) or reverberation mapping ( e.g. * ? ? ?
* ; * ? ? ?
* ) are useful techniques for weighing the smbhs at the center of galaxies and relating these to the stellar masses of their spheroids .
however , neither of these techniques can currently be applied to smgs at high redshifts and so by necessity we have to adopt much cruder methods to estimate both the black hole and spheroid masses . in this study
we use the emission arising from accretion onto the smbh as a tracer of the black hole mass and the rest - frame near - infrared luminosity of galaxy as an indicator of its stellar mass .
unfortunately , the rest - frame uv and optical spectra of the smgs provides an ambiguous tracer of the nuclear activity in these galaxies @xcite .
the difficulty of merely identifying an agn from these observations , let alone reliably estimating their luminosities is likely caused by heavy obscuration towards the smbh or faint broad lines from intrinsically weak agn being lost against the continuum of the galaxy . conversely the weakness of the agn and their obscuration makes the use of the rest - frame near - infrared luminosity a much more reliable measure of the stellar mass of the galaxies .
in contrast , hard x - ray emission from the accretion disk emerges relatively unattenuated from all but the most obscured regions , hence the x - ray emission from smgs may provide the most reliable probe of their smbh masses .
however , as demonstrated by @xcite , only the _ chandra _ 2-ms observation of the region @xcite reaches the necessary depth to provide constraints on the x - ray luminosities of typical smgs . for this reason our analysis focuses on the sample of smgs with reliable spectroscopic redshifts from @xcite which lie within this field .
we stress that precise redshifts are essential for our analysis .
not only are the mass estimates dependent on a well - constrained luminosity distance , but more critically there are strong degeneracies in redshift and age in the fitting of their spectral energy distributions ( sed ) described in 3 .
using the sample of spectroscopically identified , radio - detected smgs in the cdf - n from @xcite , obscuration - corrected x - ray luminosities were derived for 15 galaxies and smbh masses estimated assuming eddington accretion by @xcite .
we discard two of these galaxies from our analysis : j123716.01 + 620323.3 falls outside the mid - infrared coverage , and the optical / mid - infrared photometry of j123632.61 + 620800.1 is contaminated by a diffraction spike from a bright ( @xmath13 ) object only 11@xmath14 away .
these conditions prevent us from estimating the stellar masses needed in 3.2 , hence we concentrate on 13 smgs with optical , near - infrared and mid - infrared photometry or limits and reliable spectroscopic redshifts .
these galaxies range in redshift from @xmath152.9 , with a median comparable to that of the entire smg population ( @xmath0 ) .
the x - ray spectral analysis in @xcite shows that the bulk of the hard x - ray emission from the smgs in our sample is likely to be due to accretion onto the smbh , with star formation contributing only at the softest energies , @xmath16kev in the rest - frame .
the luminosity from the x - ray waveband is only a few percent of the bolometric luminosity from the smbh , and @xcite adopted a bolometric correction ( bc@xmath17 ) of @xmath18% consistent with other studies .
they also employed eddington arguments to show that @xmath19 , where the bolometric luminosity is in units of ergs@xmath10 .
the factor @xmath20 is unity for eddington - limited accretion , and @xmath21 for sub - eddington . as our bench - mark
we adopt eddington - limited accretion and use this to investigate the relation between black hole and stellar masses for our sample .
in addition to the x - ray constraints on the luminosities and hence masses of the smbhs , our analysis also requires estimates of the stellar masses of the host galaxies .
a particularly simple and relatively robust way of deriving the mass of a stellar population is to use the rest - frame near - infrared luminosity and a theoretical mass to light ratio . for all but the youngest stellar populations ,
the rest - frame near - infrared emission is dominated by light from main sequence stars and hence , for example , the rest - frame @xmath22-band luminosity changes by only a factor of two between stellar populations with ages of 100myr and 1gyr @xcite .
the rest - frame @xmath23-band luminosities of the same populations on the other hand vary by a factor of ten .
in addition to its relative insensitivity to age variations , the rest - frame @xmath22-band is over ten times less sensitive to reddening than @xmath23-band , which is important to consider for these dusty smgs . at the redshift of our targets ,
the rest frame @xmath22-band is shifted towards the mid - infrared and hence _ spitzer _ observations are required to estimate the stellar mass .
for this reason , we obtained the irac images of goods - n from the _ spitzer _ archive ( v.1 20041027 and v.1 20050505 enhanced data products ) .
we used sextractor to measure the photometry and uncertainties in these for each of the galaxies on the irac imaging . by constructing curves of growth for several isolated objects
, we determined that 6@xmath14 diameter apertures were necessary to measure the total irac flux from our targets . since source confusion is an issue for these extremely deep exposures
, we used 4@xmath14 diameter apertures and then applied a small aperture correction , determined to be 0.22 , 0.26 , 0.33 , and 0.30 mag for the 3.6 , 4.5 , 5.8 , and 8.0@xmath24 m bands respectively .
even so , we were unable to adequately constrain the 3.6@xmath24 m flux for three sources due to confusion noise from nearby sources .
we provide the irac photometry for all 13 sources in table [ tab : phot - spitzer ] . at 5.8@xmath25 m , which corresponds to a rest - frame wavelength of @xmath26 m ,
the ten high - redshift smgs at @xmath27 have a bright mean magnitude of @xmath28 , with the entire sample spanning under an order of magnitude in flux .
this is in contrast with the @xmath29-band fluxes , which cover well over an order of magnitude .
the modest dispersion in the mid - infrared fluxes of the smgs already suggests that we are dealing with a relatively homogeneous population of luminous , high redshift galaxies .
lcccc j123636.75 + 621156.1 & @xmath30 & @xmath31 & @xmath32 & @xmath33 + j123721.87 + 621035.3 & @xmath34 & @xmath35 & @xmath36 & @xmath37 + j123629.13 + 621045.8 & @xmath38 & @xmath39 & @xmath40 & @xmath41 + j123555.14 + 620901.7 & @xmath42 & @xmath43 & @xmath44 & @xmath45 + j123711.98 + 621325.7 & @xmath46 & @xmath47 & @xmath48 & @xmath49 + j123635.59 + 621424.1 & @xmath50 & @xmath51 & @xmath52 & @xmath53 + j123549.44 + 621536.8 & & @xmath54 & @xmath55 & @xmath56 + j123606.72 + 621550.7 & @xmath57 & @xmath58 & @xmath59 & @xmath60 + j123622.65 + 621629.7 & @xmath61 & @xmath62 & @xmath63 & @xmath64 + j123707.21 + 621408.1 & @xmath65 & @xmath66 & @xmath67 & @xmath68 + j123606.85 + 621021.4 & @xmath69 & @xmath70 & @xmath71 & @xmath72 + j123616.15 + 621513.7 & & @xmath73 & @xmath74 & @xmath75
+ j123712.05 + 621212.3 & & @xmath76 & @xmath77 & @xmath78 + [ tab : phot - spitzer ] to improve the constraints on the form of the seds in smgs we also make use of rest - frame uv / optical photometry to extend the wavelength coverage and so estimate the likely reddening and ages of the stellar populations . for this
we use the @xmath79 photometry presented in @xcite .
these data were supplemented with @xmath80- and @xmath81-band near - infrared imaging from the wirc2 camera at palomar ( see @xcite and @xcite for a description of these data ) .
combined , these ground - based data cover more bands than the goods - n _ hubble space telescope _ ( _ hst _ ) acs imaging of this field @xcite .
although the extra bands provided by the ground based imaging is essential for the sed template fitting ( see [ sec : stellarmass ] ) , the high - resolution imaging available from _ hst _ is valuable for studying the light profiles of these galaxies .
@xcite show that the rest - frame @xmath23-band light distributions of a sample of 27 smgs ( 5 of which are in the sample presented here ) follow a @xmath82 power law on average , consistent with the spheroidal - nature of these systems .
a summary of the optical and near - ir photometry is presented in table [ tab : phot - opt ] .
we adopt 3@xmath14 and 4@xmath14 diameter apertures to measure the fluxes from the optical and near - infrared imaging as these provide optimal signal - to - noise and a near - total estimate of the magnitudes for the smgs .
lllllllll j123636.75 + 621156.1 & @xmath83 & @xmath84 & @xmath85 & @xmath86 & @xmath87 & @xmath88 & @xmath89 & @xmath90 + j123721.87 + 621035.3 & @xmath91 & @xmath92 & @xmath93 & @xmath94 & @xmath95 & @xmath96 & & @xmath97 + j123629.13 + 621045.8 & @xmath98 & @xmath99 & @xmath100 & @xmath101 & @xmath102 & @xmath103 & @xmath104 & @xmath105 + j123555.14 + 620901.7 & @xmath106 & @xmath107 & @xmath108 & @xmath109 & @xmath110 & @xmath111 & @xmath112 & @xmath113 + j123711.98 + 621325.7 & @xmath114 & @xmath115 & @xmath116 & @xmath117 & @xmath118 & @xmath119 & @xmath120 & @xmath121 + j123635.59 + 621424.1 & @xmath122 & @xmath123 & @xmath124 & @xmath125 & @xmath126 & @xmath127 & @xmath112 & @xmath128 + j123549.44 + 621536.8 & @xmath129 & @xmath130 & @xmath131 & @xmath132 & @xmath133 & @xmath134 & @xmath135 & @xmath136 + j123606.72 + 621550.7 & @xmath137 & @xmath138 & @xmath139 & @xmath140 & @xmath141 & @xmath142 & @xmath120 & @xmath143 + j123622.65 + 621629.7 & @xmath144 & @xmath145 & @xmath146 & @xmath147 & @xmath148 & @xmath149 & @xmath120 & @xmath150 + j123707.21 + 621408.1 & @xmath151 & @xmath152 & @xmath153 & @xmath154 & @xmath155 & @xmath156 & @xmath120 & @xmath157 + j123606.85 + 621021.4 & @xmath158 & @xmath159 & @xmath160 & @xmath161 & @xmath162 & @xmath163 & @xmath164 & @xmath165 + j123616.15 + 621513.7 & @xmath151 & @xmath166 & @xmath167 & @xmath168 & @xmath169 & @xmath170 & @xmath120 & @xmath171 + j123712.05
+ 621212.3 & @xmath151 & @xmath172 & @xmath172 & @xmath173 & @xmath174 & @xmath175 & @xmath120 & @xmath176 + [ tab : phot - opt ] we plot the combined optical , near- and mid - infrared photometry for the 13 smgs in fig .
[ fig : allseds ] .
the seds for these galaxies show three clear features : a sharp decline at rest - frame wavelength of @xmath177 , resulting from the lyman@xmath178 limit within the galaxy and the foreground intergalactic medium ; a break at around 4000 arising either from the spectral breaks found in young or evolved stellar populations at this approximate wavelength and finally a peak in the emission at @xmath179 m associated with the minimum in the h@xmath180 opacity and hence the maximum emission from stellar atmospheres .
looking at the seds in fig .
[ fig : allseds ] we conclude that there is a modest range in the rest - frame uv / optical / near - infrared properties of radio - detected smgs .
however , focusing on the ten higher - redshift smgs , at @xmath27 , of those with full coverage of their rest - frame uv and optical seds , 80% show a detectable discontinuity in their seds around 4000 in the rest - frame ( fig . [
fig : allseds ] ) .
we interpret this break as arising from the balmer limit in stellar populations with ages of @xmath181 s myrs .
this interpretation is consistent with the likely nature of smgs as luminous far - infrared sources seen during a phase of intense star formation .
we note that age - related differential reddening may also account for the relative strength of this feature in an even younger on - going starburst @xcite .
due to the complex mix of extinction expected within smgs @xcite , which may influence different age stellar populations to different degrees , we have adopted a very simple approach in determining our stellar masses .
we first estimate the rest - frame @xmath22-band luminosities of the galaxies , interpolated if necessary from a fit to the sed , and then combine these with estimates of the mass - to - light ratios in the @xmath22-band for a range of plausible ages for the dominant stellar population . to derive the rest - frame @xmath22-band luminosities we exploit the hyper - z package to fit physically meaningful spectral templates to the spectral energy distributions of each galaxy and then estimate their @xmath22-band luminosities by interpolation of the best - fit templates .
this approach has the added advantage that the properties of the best - fit spectral templates may provide some insights into the properties of the stellar populations within these galaxies .
we restrict the code to fit the seds at the spectroscopically confirmed redshift , and allow it to choose between two model sed templates : one having an instantaneous burst of star - formation , or one with a constant star formation ( csf ) history .
these represent the two extremes of the likely star formation histories for these young galaxies @xcite and provide an adequate variety to guarantee a reliable estimate of the @xmath22-band luminosity for those galaxies whose seds are dominated by star - light . in our fits ,
the reddening ( @xmath182 ) was free to vary , and the ages at the galaxy s redshift are constrained to be physical within our adopted cosmology .
these provide additional degrees of freedom for the fitting and may give some indication of the properties of the composite stellar population in these systems . for each galaxy ,
hyper - z returns the absolute magnitude in the rest - frame @xmath22-band , as well as the age , reddening and star formation history of the best - fit template .
although the m@xmath183 values are not corrected for it , the reddening ( which is @xmath184a@xmath185 based on the extinction law presented in @xcite ) has a negligible effect on our stellar mass estimates , particularly compared with other uncertainties in our analysis .
the typical fit to a sed in our sample produces a reasonable @xmath186 , with only three minor exceptions : both j123635.59 + 621424.1 and j123606.72 + 621550.7 are spectroscopically identified agn , and demonstrate an obvious excess at the reddest wavelengths .
this would be expected , since the input templates do not account for the contribution in the mid - ir of warm dust from an agn .
j123549.44 + 621536.8 also has a moderate near - ir excess , but the significance is muted somewhat by its proximity to the edge of the irac map , where photometry was more difficult to estimate .
the starburst j123606.85 + 621021.4 has an excess in the uv , but this is caused by a contribution from an unrelated blue foreground @xmath187 galaxy only 2@xmath14 away . in all these cases , we feel the best - fit sed template is still a good measure of the rest - frame @xmath22-band luminosity and provides a good estimate of the underlying stellar mass . as we show later , changes to the implied stellar masses of factors of 2 ( due to agn contamination at restframe @xmath22-band ) , would not significantly alter the conclusions of this work . to convert these estimates for m@xmath183 into stellar masses we need to determine the light - to - mass ratio , l@xmath183/m for the dominant stellar populations in these galaxies . to do this
we turn to the starburst99 stellar population model @xcite , which provides estimates of m@xmath183 for our two extreme scenarios either a burst or constant star formation history .
we adopt the the initial mass function ( imf ) of @xcite , this is also the same as that used for the hyper - z templates .
the imf has a mass range of 0.1125m@xmath2 and assumes solar metallicity .
higher metallicities are often favored when fitting the seds of ultraluminous infrared galaxies ( ulirgs ) in the local universe ( e.g. * ? ? ?
* ) , but we found no appreciable difference in the estimates of l@xmath183/m for a range of likely values . we can combine the predicted @xmath22-band luminosities as a function of time from starburst99 with the integrated stellar masses at that time to derive the effective light - to - mass ratio for the stellar population : @xmath188 where 3.3 is the absolute magnitude of the sun in the @xmath22-band @xcite .
we find that the light - to - mass ratio for the csf case is reasonably described by a simple power law , for @xmath189myr .
the behavior of l@xmath183/m for a burst model is not as easy to parameterize over the likely range of ages of smgs , so instead we use tabulated values . in solar units , we find that l@xmath183/m is [ 2.5 , 2.2 , 13.3 , 7.6 , 4.2 , 2.7 , 1.9 ] at 10 , 20 , 50 , 100 , 200 , 500 , 1000myrs respectively .
we caution the reader that at ages of @xmath190myrs changes to the adopted imf or metallicity can have significant effects on the l@xmath183/m and so the predictions for extremely young bursts are much more uncertain . in principle
, we can constrain the ages of the dominant stellar population in each smg from our sed modeling .
however , the correlated nature of the fits and the fact that we are likely sampling different mixes of stellar populations at different wavelengths due to differential reddening means that the results are highly uncertain for individual galaxies . to demonstrate this point
, we note that all of the galaxies can be reasonably fit by _ either _ model star formation history , but with very different ages .
this is a direct consequence of the fact that an older galaxy with a constant star - formation history has a similar sed to that of a younger burst model , a point also noted in @xcite .
for this reason we prefer to use only ensemble averages for the derived parameters .
if we restrict our hyper - z fits to the burst model only , we estimate ensemble averages of @xmath191myr , @xmath192 and m@xmath193 . alternatively , restricting our fits to just the continuous star - forming model yields @xmath194myr , @xmath195 and m@xmath196 . for the average age from our fits of the burst model
, we estimate a typical l@xmath183/m=3.7 , while the average for the continuous star - forming model is l@xmath183/m=2.6 .
hence , the average stellar mass from the samples , estimated using @xmath197 and l@xmath183/m , are essentially identical , @xmath198 and @xmath199 . based on these results ,
we choose to simply use l@xmath183/m=3.2 for all galaxies .
given the degeneracy between the sed fits , we use the average m@xmath183 from the burst and csf fit for each galaxy , and half the difference between them as a measure of the uncertainty .
we summarize the results of our stellar mass estimates from the sed fitting in table [ tab : hyperz ] .
the mean stellar mass is @xmath200 for the sample of smgs used here .
llllccl j123636.75 + 621156.1 & 0.557 & @xmath201 & @xmath202 & 42.7 & @xmath203 & + j123721.87 + 621035.3 & 0.979 & @xmath204 & @xmath205 & 43.7 & @xmath206 & + j123629.13 + 621045.8 & 1.013 & @xmath207 & @xmath208 & 43.2 & @xmath209 & + j123555.14 + 620901.7 & 1.875 & @xmath210 & @xmath211 & 44.4 & @xmath212 & + j123711.98 + 621325.7 & 1.996 & @xmath213 & @xmath214 & 43.6 & @xmath206 & + j123635.59 + 621424.1 & 2.005 & @xmath215 & @xmath216 & 44.0 & @xmath217 & near - ir excess + j123549.44 + 621536.8 & 2.203 & @xmath218 & @xmath219 & 44.0 & @xmath220 & near - ir excess + j123606.72 + 621550.7 & 2.416 & @xmath221 & @xmath222 & 43.8 & @xmath223 & near - ir excess + j123622.65 + 621629.7 & 2.466 & @xmath224 & @xmath225 & 44.0 & @xmath226 & + j123707.21 + 621408.1 & 2.484 & @xmath227 & @xmath228 & 43.8 & @xmath229 & + j123606.85 + 621021.4 & 2.505 & @xmath230 & @xmath231 & 43.7 & @xmath206 & uv excess + j123616.15 + 621513.7 & 2.578 & @xmath232 & @xmath233 & 43.7 & @xmath206 & + j123712.05 + 621212.3 & 2.914 & @xmath234 & @xmath235 & 43.4 & @xmath236 & + [ tab : hyperz ] fig .
[ fig : m - lx ] shows the rest - frame @xmath23-band and @xmath22-band luminosities from our sed fits against the x - ray luminosities from @xcite .
there is an apparent trend between @xmath22-band and x - ray luminosity , which is reasonably fit by a power - law of the form l@xmath239l@xmath240 . to test the strength of the correlation
, we applied a spearman rank test , the correlation coefficient is 0.68 , with a 2.4@xmath241 deviation from the null - hypothesis that l@xmath183 and l@xmath17 are not correlated . removing the three lower redshift sources in fig .
[ fig : m - lx ] does weaken the apparent correlation : the power law index becomes @xmath242 , while the spearman test estimates a 1.8@xmath241 deviation from the null - hypothesis .
while the modest dynamic range of our sample makes evidence for a correlation difficult to find , as we discuss below , it does yield strong constraints on the normalization of the relationship .
for both the full sample and with the three lower redshift sources excluded , the trends with @xmath22-band magnitude are three times tighter than that seen between the reddening - corrected rest - frame @xmath23-band magnitude and x - ray luminosities , yielding an rms of 0.6 versus 2.0mags respectively .
this difference most likely reflects the increased sensitivity to reddening corrections in the rest - frame @xmath23-band , in addition to the fact that near - infrared luminosities better probe the evolved stellar population .
since the uncertainty in the stellar mass estimates scale as @xmath243 dex , these results underline the advantages of the rest - frame @xmath22-band luminosities to estimate the stellar masses of these galaxies .
if we assume the x - ray emission arises from accretion onto smbhs , as indicated by the spectral analysis of @xcite , we can interpret the correlation between @xmath244 and @xmath245 as one between the stellar component of the galaxies and their smbhs .
such a correlation would also be expected if the @xmath22-band light were due to hot dust powered by the agn . however , with only a few exceptions , the seds seem to be dominated by stellar light , as evident from a well - defined stellar bump peaking near 1.6@xmath24 m ( fig . [
fig : allseds ] ) . for the remainder of the paper
we will assume that the @xmath22-band is tracing the stellar mass , but the reader should bear the caveat in mind . in fig .
[ fig : magorrian ] we plot the distribution of m@xmath237 the stellar mass from the sed fits assuming a fixed l@xmath183/m , and m@xmath246 the smbh masses derived from the absorption - corrected x - ray luminosities by @xcite assuming eddington accretion and a @xmath247% bolometric correction . as in fig .
[ fig : m - lx ] , we see a reasonable correlation between these two quantities .
the limited dynamic range of our sample and the uncertainties in the individual measurements means we can not provide a reliable constraint on the power - law index for this correlation , m@xmath248m@xmath249 .
the data are consistent with @xmath2500.31 with a very large uncertainty depending upon the exact choice of sample used
. nevertheless , while the slope of the correlation is uncertain , the normalization is well - constrained at the mean m@xmath237 of our sample . at @xmath251 ,
we find that the average eddington - limited black - hole mass is @xmath252m@xmath253 .
if we exclude the three smgs at @xmath254 , the mean stellar and black hole masses are slightly larger : @xmath255 and @xmath252m@xmath256 as a comparison we turn to a sample of nearby galaxies with reliable black - hole masses estimated from high resolution _ hst
_ spectroscopy , and stellar masses derived via near - infrared observations of the host galaxies @xcite .
although their full sample includes 37 galaxies , we concentrate on the 27 `` group 1 '' systems judged by those authors to be the most secure . using these data
, @xcite showed that : @xmath257 we note that this relation predicts black hole masses roughly 50% higher than those presented in @xcite , who used jeans equation modeling rather than virial estimators to predict the bulge mass .
using these relations , our data suggests that the black hole masses we infer for smgs ( either in the full sample or for the ten smgs at @xmath27 ) are @xmath258 times smaller than for similarly massive local spheroids .
with the assumptions about the l@xmath183/m and eddington fraction @xmath20 we have adopted , our results suggest that the black holes in smgs require significant growth in order to approach the present - day relation . to understand the significance of this conclusion ,
however , we need to better characterize any biases in the mass estimates for both of these components .
note that errors on individual sources are large in principle , since for both the black - hole and stellar masses we have assumed ensemble values ( i.e. fixed @xmath20 for the black - holes , and a single l@xmath183/m value for all stellar mass estimates ) .
it is quite likely that source - to - source variations will affect the scatter of the data , but until more accurate methods are available for the mass estimates , we concentrate instead on understanding any systematic bias which could effect the sample as a whole . altogether , the ratio between between our black hole masses and those in the nearby universe scales as : @xmath259 where we have simplified the expression by assuming the power - law relation between stellar and black hole has an exponent of 1.0 .
in fig . [
fig : range ] we plot tracks of different conditions where the above expression is unity .
this demonstrates how difficult it is for any one factor to explain the difference between our sample and the local relation , and suggests that either the discrepancy is real , or that biases in the parameters must conspire if , in fact , smgs have a black - hole / stellar mass ratio comparable to local systems .
we now discuss these potential biases in more detail .
we begin by pointing out that the stellar masses we derived are on average five times higher than those reported from modeling the uv / optical seds of smgs in @xcite .
this contrasts with the reasonable agreement in the estimated stellar masses derived for uv - selected galaxies at similar redshifts to the smgs , by shapley et al .
( 2005 ) from both rest - frame uv / optical and rest - frame uv / optical / near - infrared photometry .
we attribute the difference in our masses to those from @xcite to the absence of mid - infrared information to constrain the sed modeling in the earlier work , which is then much more sensitive to the reddening corrections derived from the sed fits to these very dusty galaxies and the possibilities of strong differential reddening within their young stellar populations . in our modeling of the stellar mass of the systems we adopted a l@xmath183/m ratio of 3.2 , characteristic of a burst with an age of @xmath260myrs .
if we require that the smg and the present - day m@xmath237m@xmath238 relations agree , then from eq .
[ equ : unity ] we see that the stellar masses must drop by a factor of @xmath258 .
this corresponds to l@xmath183/m@xmath261 , which is not possible with the miller - scalo imf . for a salpeter imf ,
such a high l@xmath183/m ratio implies ages for the smgs of @xmath262myrs .
we discuss the implications for varying the imf later in this section .
however , we note that there are fundamental inconsistencies in our calculations if we rigidly interpret the star formation histories we used .
the smgs in our sample have been selected on the basis of their intense far - infrared luminosities , which implies high star formation rates .
the fast decline in the far - infrared emission from a starburst after star formation ceases indicates that star formation is currently on - going in these systems or has declined significantly in only the last few 10 s of myrs .
the submm detection of these sources is therefore formally inconsistent with the star formation history in our burst model where there has been no star formation within the galaxy in the past @xmath261myrs .
although the alternative csf models include on - going star formation , the much higher ages derived from these fits are equally inconsistent with a constant rate of star formation at the level indicated by the observed far - infrared emission from the smgs , @xmath812@xmath263m@xmath2yr@xmath10 . to form an @xmath264 galaxy with a constant sfr at this level
would take just @xmath8100myrs , not the 2gyrs derived from the fits .
there are two solutions to solve these inconsistencies within our models : 1 ) allow a more complex star formation history , most likely involving a recent burst of activity ; 2 ) apply age - dependent dust extinction to the stellar populations , with the youngest stars being most extincted .
as both solutions involve additional free - parameters to what is already a weakly - constrained problem , we will only qualitatively discuss their impact .
bursts of star formation are a natural consequence of interaction - induced activity due to the cyclic behavior of bound orbits @xcite .
they lead to a dispersion in the properties of smgs dependent on the number of previous bursts they have experienced .
the main consequence in terms of our analysis is the realization that the present activity may not have produced the total stellar population we observe .
mixing in a fraction of older stars will systematically decrease the effective l@xmath183/m , potentially increasing the masses of the galaxies .
this would therefore push the smgs even further away from the @xmath265 m@xmath237m@xmath238 relation .
age - dependent or selective extinction of the stellar populations ( e.g. * ? ? ?
* ) within smgs can preferentially extinguish the uv emission from the youngest stars .
this results in seds with much redder spectral indices between the uv and near - infrared than can be achieved with a simple foreground dust screen acting upon the whole stellar population .
when the diffusion timescale for stars to escape their obscured birth - places is around a few 100 s myrs , this process can also help reproduce the strength of the spectral breaks at @xmath84000 , even in galaxies with significant on - going star formation . as a result ,
selective extinction would provide the opportunity to fit the sed of our smg sample with relatively young csf models , @xmath8100myrs , consistent with the ages needed to produce their observed near - infrared luminosities given the star formation rates inferred from their far - infrared emission .
attempts to estimate typical ages for smgs include the spectral dating of the few examples with high - quality uv spectroscopy .
such an analysis by @xcite indicates an age for a uv - bright smg of @xmath266myrs , similar to the minimum age we require .
however , as @xcite state the smg they model may be a rare system and the small proportion of smgs showing similar signs of young stellar populations in their spectra suggests a more typical age for an smg would be 50100myrs .
another estimate of the likely age of smgs can be derived via gas - depletion timescales using the measured mass of the gas reservoir and the estimated star formation rates suggest likely lifetimes of @xmath267myrs @xcite , where we have assumed an agn contribution to the far - infrared luminosities in line with @xcite . finally , taking our estimated stellar masses and the median star formation rates derived from the far - infrared luminosities of the smgs in our sample using our adopted imf
, we determine lifetimes of @xmath181myrs to produce the stellar populations we see . balancing these estimates with those from the sed fitting ,
we conclude that the typical ages of smgs are likely to be @xmath268myrs .
hence the stellar masses are inconsistent with what would be predicted from the local m@xmath237m@xmath238 relation assuming our black - hole masses are accurate .
are there other ways to change the effective l@xmath183/m without invoking such young ages ? one way to reduce it is to have a significant contribution to the rest - frame @xmath22-band light from the agn .
however the bulk of our sample , as evident in fig .
[ fig : allseds ] , are well - fit by a stellar component only . in the few cases where agn cause the sed fits to diverge at the longer wavelengths ,
the estimated luminosity we derive at 2.2@xmath25 m is still dominated by the stars .
we used the 8@xmath24 m excess ( the ratio of the measured 8@xmath24 m flux over that derived from the best fit model template ) to quantify possible agn contamination .
we find that the @xmath27 sources ( excluding the 3 marked as having an obvious near - ir excess ) have an average excess of 1.0 , suggesting little or no contribution from an agn .
the remaining three sources however , have an average excess of @xmath269 , which means a nearly equal contribution to the rest - frame @xmath22-band from agn as stars . scaling down the stellar masses of these three galaxies to account for this results in only a @xmath184 reduction in the mean stellar mass of the sample , which is small considering the other factors we discuss .
in addition to contribution of light from the agn , we note that our analysis includes light from _ all _ of the stars , and not simply the bulge component used in deriving the local relation . while the high - resolution _ hst_/nicmos imaging verifies that the rest - frame @xmath23-band light in smgs are reasonably fit by a @xmath82 surface brightness profile @xcite
, some systems do show a significant disk component .
whether these stars are still in a disk component at the present - day or whether they would have been driven into the bulge through , for example , bar - instabilities , is impossible to verify , and so we flag this to the reader as a fundamental uncertainty of our analysis .
nevertheless , we note that any correction to the stellar masses due to a disk component is unlikely to change the inferred spheroid mass by more than a factor of 2@xmath270 for a small fraction of the smgs .
the imf also has a significant effect on stellar mass estimates .
@xcite argue that an imf biased to more massive stars is required to reproduce the properties of smgs in current semi - analytic galaxy formation models , and invoke a power - law imf with an index flatter than the canonical salpeter @xcite slope of @xmath271 .
we used starburst99 to study the effect of this top - heavy imf on our mass estimates , but find that the stellar mass estimates change by at most a factor of @xmath272 for the range of likely ages for smgs .
an additional factor of three in the light - to - mass ratio can be picked up by truncating the imf at @xmath273m@xmath2 , and there are different imf prescriptions ( such as the one proposed by @xcite that has an increased @xmath22-band contribution from thermally pulsating agb stars ) that can also increase the effective light - to - mass ratio . altogether it is possible that these effects could conspire to decrease our stellar masses by up to a factor of @xmath274 .
however , we conclude that any bias present in the stellar mass is insufficient by itself to account for the difference between our results and the local relation .
are the high stellar masses we derive in conflict with other estimates of the masses of smgs ?
dynamical masses for smgs from co spectroscopy with millimeter interferometers indicate a median gas+stellar masses of @xmath275 within the central @xmath262kpc of typical smgs with 850-@xmath25 m fluxes of a few mjys @xcite .
in addition , five of the galaxies in our sample have dynamical mass estimates from h@xmath276 spectroscopy @xcite , which all give constraints on the masses of @xmath27711.3 , again on @xmath278kpc scales . in a recent analysis , @xcite use the integrated co line luminosities for 12 smgs to estimate an average contribution from gas of @xmath279 within @xmath262kpc for these systems . thus taking into account the contribution from gas ,
the dynamical estimates are consistent with the presence of @xmath280 of stars within the central regions of the smgs . as highlighted by @xcite the h@xmath276 dynamical mass estimates derived from the h@xmath276 line widths for smgs are five times higher than similar constraints on uv - selected galaxies at comparable redshifts @xcite .
to see if this difference is also reflected in the stellar mass estimates for the two populations , we ran our hyper - z analysis on the sample of 72 spectroscopically confirmed , uv - selected , and irac - detected , galaxies at @xmath281 presented in @xcite .
these uv - selected galaxies @xcite make an ideal comparison set , as they are more modestly star - forming galaxies at roughly the median redshift of our sample . as with our analysis , the fits to the data from @xcite were degenerate in sed template , age and reddening .
roughly one - third of the objects are best fit by a young ( @xmath282myr ) burst while the remainder are better modeled by a constant star - formation history with age of @xmath283myr .
again we find that the choice of model has little impact on the derivation of the stellar mass , and we obtain a mean of @xmath284 for this sample .
the estimated ages and masses are in good agreement to the results which @xcite derive using a more detailed analysis ( @xmath285 ) .
this confirms that our approach to stellar mass estimation is robust and also suggests that the smgs in our sample are on average @xmath274 times more massive than the uv - selected population at @xmath286 .
overall , we conclude that our stellar masses are consistent with dynamical limits on the masses of the smg population and support the suggestion that the central regions of these galaxies are baryon dominated .
we note that the co detections of smgs indicate that there is sufficient gas present to continue to build the black hole , but at the same time the gas masses are small enough that they will not appreciably change the stellar masses even if the bulk of it is converted into stars . only by bringing in more gas or stars into these systems ( through cooling or mergers )
can their stellar masses be significantly increased from what we measure .
is it possible that the sources only appear to lie below the local relation because the black - hole mass estimates from @xcite are too low ? certainly , in the light of the discussion above it appears that the black hole mass estimates are probably the more important source of uncertainty in our analysis . in estimating the black - hole masses , @xcite assumed eddington - limited accretion and a @xmath18% bolometric correction to convert the observed absorption - corrected l@xmath17 to m@xmath246 . since the eddington limit appears to provide a reasonable upper bound to the accretion rate of black holes ( e.g. , *
* ) , this approach essentially provides a lower limit on the black - hole masses and sub - eddington accretion would imply larger black holes .
however , an average black - hole mass @xmath258 times higher is needed to reconcile our data at the characteristic stellar mass of the smgs , @xmath1m@xmath2 .
this would require @xmath287 , which is low given the gas - rich environments in which the smbhs are likely to reside .
although we do not know the eddington rate of these galaxies , the large fraction of smgs that host agn activity suggests that the accretion is likely to be reasonably efficient ( see @xcite ) . there is tangential support for this claim from studies that have estimated black - hole masses for other active high - redshift populations independently of the eddington rate , finding @xmath2880.11.7 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . at the smallest accretion efficiencies found in these studies , the black - hole masses would still be @xmath85 times too small for our estimated stellar masses .
there is further support for eddington - limited or near eddington - limited accretion from the similarity between the optical / near - infrared spectra of some smgs with those of local narrow line seyfert 1 ( nls1 ) galaxies ; see 4.3 of @xcite . if the spectral similarities of the agn signatures in these two populations are suggestive of comparable physical conditions in the accretion disk around the smbh , then the approximately eddington - limited rates estimated for local nls1s may be appropriate for smgs .
moreover , nls1s appear to have smaller black holes than those predicted from the local m@xmath237m@xmath238 relation ; using reverberation mapping , @xcite and @xcite show that that the smbhs in nls1s are an order of magnitude smaller than predicted for galaxies having a stellar mass of @xmath289m@xmath2 .
thus physical models for the accretion disks and smbhs in nls1s may be very relevant for understanding the growth of smbhs in massive , young galaxies at high redshifts .
we do caution the reader that the broad - line region in smgs may be missed due to heavy extinction in these dusty galaxies , and hence the spectra may be dominated by a less obscured narrow - line region .
high signal - to - noise spectra would be useful to search for faint , broad wings that may be otherwise missed .
there is also theoretical support for @xmath290 efficiencies : @xcite and @xcite predict that this condition will be met near the peak epoch of star formation , where these objects have a stronger chance of being detected at submm wavelengths .
interestingly , these models also predict black hole masses @xmath291 12 orders of magnitude smaller than the local relation during the obscured , star - bursting phase , similar to what our data suggest .
hence , while the assumption of eddington limited accretion is a large one , it is not a unreasonable choice .
we reiterate that even an efficiency 10 times smaller would still result in the black - holes in smgs being smaller than would be predicted by the local relation .
another uncertainty in the smbh estimates comes from the bolometric correction applied to the x - ray luminosities .
@xcite assumed that @xmath18% of the energy output due to accretion was emitted in the x - ray waveband @xcite , but if this was lower , the inferred mass of the smbh would rise .
however , they also pointed out that this correction is similar to the estimated bolometric corrections for obscured and unobscured agn of similar luminosity to the agns in the smgs ( see * ? ? ? * ) .
a final potential source of uncertainty in the smbh masses comes from the possibility that a substantial fraction of the x - ray emission from the smbh is still obscured at rest - frame energies of @xmath820kev .
for instance , @xcite find that the _ chandra _
2-ms exposure of the cdf - n fails to recover approximately 40% of the hard x - ray background .
this may in part be hidden within highly - obscured ` type-2 ' quasars @xcite , but dusty smgs also harbor obscured agn ( and could in some cases be type-2 quasars ) .
indeed , not all smgs are detected in the _ chandra _
2-ms x - ray image and so potentially some fraction of the x - ray emission from the smbh in this population could be missed from the _ chandra _ analysis . in 3.5 of @xcite ,
it is pointed out that while obtaining tight constraints on the obscuring column densities for individual sources is difficult , the agreement between their models and x - ray spectra of smgs ranked by obscuration class is excellent , suggesting that most of the x - ray emission in these sources is accounted for .
if a large fraction of the x - ray emission in smgs is still absorbed , we will need to wait for the next generation of x - ray observatories to find it . to summarize the previous two subsections
, it is possible to find combinations of the parameters used to estimate the stellar and black hole masses of smgs which would reconcile them with the local m@xmath237m@xmath238 relation .
for example , if we assume a 6% bolometric correction , then we require a combination of a typical smg age of just @xmath292myrs and an eddington efficiency of @xmath293 to place the high - redshift data on the local relation ( fig .
[ fig : range ] ) .
this combination is plausible , and if it is the case then the implication is that the stellar and black hole mass grow simultaneously , with a mass ratio that is similar at high redshift as well as low .
however , the need for both low eddington accretion and a very short lifetime for the smgs is uncomfortable and hence we should also explore the possibility that the black holes in smgs are , in fact less massive than expected from their stellar masses and the low redshift m@xmath237m@xmath238 relation .
the following discussion assumes that this is indeed the case .
the theoretical models of @xcite suggest that smbhs may grow exponentially , taking only a few 10 s of myr to mature @xcite .
this is a short enough time - scale that smgs with more evolved smbhs , which lie on the local relationship , should fall within the redshift range probed by our sample . that we do not see such systems
implies that as the black hole rapidly builds up its mass , it either shuts down the star - formation that powers the far - infrared luminosities ( hence they drop out of the submillimeter samples ) or they become so obscured that they are no longer detected in the x - ray waveband .
the former argument supports a scenario that links the qso and smg populations .
as the black hole undergoes a rapid buildup of mass , the ensuing strong radiation field strips away the obscuring gas and dust , terminating star formation and eventually allowing the agn to be visible for a short time as a luminous optical quasar .
the close similarity between the redshift distribution of radio - identified smgs and qsos @xcite also lends support to this hypothesis , as does the relative space densities of smgs and qsos given our estimates of their respective lifetimes @xcite .
in addition , recent high resolution adaptive optics imaging by @xcite has shown that the black hole masses and host galaxy luminosities of @xmath294 quasars are consistent with the local m@xmath295m@xmath238 relation , suggesting that qsos could be the end - product of the rapid growth of smbhs seen soon after an smg - like phase .
finally , work by @xcite using scuba to measure the rest - frame far - infrared luminosities of unobscured and moderately obscured qsos at @xmath296 verify that the former class represent a phase of significant smbh growth but not bulge growth , while the latter may be a transition phase between the star - formation dominated smgs and agn - dominated quasars .
thus , if smgs are the progenitors of quasars , they _ must _ lie below the locally determined m@xmath237m@xmath238 relation . in conclusion
, there is good qualitative and quantitative agreement for an evolutionary sequence which connects moderate mass smbhs in far - infrared luminous , star - forming galaxies at @xmath2963 ( when most of the spheroid is produced ) , with subsequent growth of the smbhs during a short phase as a luminous quasar to result in massive spheroids at the present - day which lie on the m@xmath237m@xmath238 relation .
using deep _ chandra _ , _ spitzer _ , and ground - based optical and near - infrared imaging data , we have estimated the stellar and black hole masses for a sample of 13 smgs with precise spectroscopic redshifts .
our main conclusions are : 1 . there is a tentative correlation between the rest - frame near - infrared and x - ray luminosity in this sample of smgs .
the x - ray emission likely arises from accretion of matter onto a smbh , and so can be used to estimate its mass .
we find a similar correlation between rest - frame x - ray and @xmath23-band luminosities , but this exhibits more scatter , suggesting that it is more sensitive to dust and age variations in these young , obscured galaxies . 2 .
the restframe uv optical near - infrared seds of most of our smgs are dominated by stellar emission and not light from the agn , and are well - described by either a young burst model , or by an older galaxy with a more continuous star - formation history . in both cases , the derived light - to - mass ratios , and hence stellar mass estimates , are similar .
the mean stellar mass for the sample is @xmath251 , consistent with h@xmath276 and co dynamical mass estimates for the smg population .
hence the _ spitzer_observations have shown that these intense star - forming galaxies have a significant stellar population in place at @xmath286 .
we find that smgs are typically 10 times more massive than typical uv - selected galaxies at similar redshifts .
we use the assumption of eddington - limited accretion to estimate the black hole masses for each galaxy in the sample .
we find that for a given stellar mass , the black holes are @xmath258 times less massive than those for spheroids with similar stellar masses in the local universe .
6 . in order to determine if the high - redshift smgs lie of the local m@xmath237m@xmath238 relation , we explore the range of assumptions that are used to estimate both the stellar and black hole masses .
we find that most reasonable combinations of parameters lead to lower black hole masses at a fixed stellar mass in smgs at @xmath294 compared to the low redshift relation for spheroids .
we suggest the nls1 galaxies may be a good template for understanding the growth of smbhs at high redshift .
they have spectral properties common with many smgs , and ( locally ) have black hole masses smaller than predicted for less active spheroids .
our results are consistent with a model where the smbhs in smgs subsequently undergo rapid growth to reach the local m@xmath237m@xmath238 relation .
such a phase would most naturally be explained if the smgs evolved into quasars on their way to becoming luminous , passive ellipticals at the present - day .
we thank christine done , duncan farrah and rowena malbon for useful conversations , and mark brodwin for advice on irac photometry measurements .
we also thank an anonymous referee for a constructive report which helped to improve the structure and content of this paper . irs and
dma acknowledge support from the royal society .
awb is supported by nsf ast-0208527 , the sloan foundation and the research corporation .
adelberger , k. l. , steidel , c. c. , shapley , a. e. , hunt , m. p. , erb , d. k. , reddy , n. a. , & pettini , m. 2004 , , 607 , 226 alexander , d. m. , et al .
2003 , , 125 , 383 alexander , d. m. , smail , i. , bauer , f. e. , chapman , s. c. , blain , a. w. , brandt , w. n. , ivison , r. j. 2005a , , accepted alexander , d. m. , smail , i. , bauer , f. e. , chapman , s. c. , blain , a. w. , brandt , w. n. , ivison , r. j. 2005b , , accepted barger , a. j. , cowie , l. l. , mushotzky , r. f. , yang , y. , wang , w .- h . , steffen , a. t. , & capak , p. 2005
, , 129 , 578 barger , a. j. , cowie , l. l. , & richards , e. a. 2000 , , 119 , 2092 baugh , c. m. , lacey , c. g. , frenk , c. s. , granato , g. l. , silva , l. , bressan , a. , benson , a. j. , & cole , s. 2005 , , 356 , 1191 benson , a. j. , bower , r. g. , frenk , c. s. , lacey , c. g. , baugh , c. m. , & cole , s. 2003 , , 599 , 38 bertin , e. , & arnouts , s. 1996 , , 117 , 393 blain , a. w. , chapman , s. c. , smail , i. , & ivison , r. 2004 , , 611 , 725 bolzonella , m. , miralles , j .-
m . , & pell ' o , r. 2000 , , 363 , 476 boroson , t. a. 2002 , , 565 , 78 borys , c. , scott , d. , chapman , s. , halpern , m. , nandra , k. , & pope , a. 2004 , , 355 , 485 borys , c. , chapman , s. , smail , i. r. , & ivison , r. j. 2006 , , in prep .
botte , v. , ciroi , s. , rafanelli , p. , & di mille , f. 2004 , , 127 , 3168 bundy , k. , ellis , r. s. , & conselice , c. j. 2005 , , accepted calzetti , d. , armus , l. , bohlin , r. c. , kinney , a. l. , koornneef , j. , & storchi - bergmann , t. 2000 , , 533 , 682 capak , p. , et al . 2004 , , 127 , 180 cavaliere , a. , & vittorini , v. 2002 , , 570 , 114 chapman , s. c. , smail , i. , blain , a. w. , & ivison , r. j. 2005 , , accepted chapman , s. c. , smail , i. , windhorst , r. , muxlow , t. , & ivison , r. j. 2004 , , 611 , 732 chapman , s. c. , blain , a. w. , ivison , r. j. , & smail , i. r. 2003 , , 422 , 695 collin , s. , & kawaguchi , t. 2004 , , 426 , 797 cox , a. n. 2000 , allen s astrophysical quantities , 4th ed . publisher : new york : aip press ; springer , 2000 . edited by by arthur n. cox .
isbn : 0387987460 dawson , s. , mccrady , n. , stern , d. , eckart , m. e. , spinrad , h. , liu , m. c. , & graham , j. r. 2003 , , 125 , 1236 dickinson , m , et al .
2005 , in preparation di matteo , t. , springel , v. , & hernquist , l. 2005 , , 433 , 604 elvis , m. , et al .
1994 , , 95 , 1 erb , d. k. , shapley , a. e. , steidel , c. c. , pettini , m. , adelberger , k. l. , hunt , m. p. , moorwood , a. f. m. , & cuby , j. 2003 , , 591 , 101 fabian , a. c. 1999 , , 308 , l39 farrah , d. , surace , j. a. , veilleux , s. , sanders , d. b. , & vacca , w. d. 2005 , , accepted frayer , d. t. , et al . 1999 , , 514 , l13 frayer , d. t. , ivison , r. j. , scoville , n. z. , yun , m. , evans , a. s. , smail , i. , blain , a. w. , & kneib , j .-
1998 , , 506 , l7 gebhardt , k. , et al .
2000 , , 543 , l5 giavalisco , m. , et al .
2004 , , 600 , l93 granato , g. l. , de zotti , g. , silva , l. , bressan , a. , & danese , l. 2004 , , 600 , 580 greve , t. r. , ivison , r. j. , bertoldi , f. , stevens , j. a. , dunlop , j. s. , lutz , d. , & carilli , c. l. 2004 , , 354 , 779 greve , t. r. , et al .
2005 , , accepted grupe , d. , & mathur , s. 2004 , , 606 , l41 h " aring , n. , & rix , h. 2004 , , 604 , l89 ivison , r. j. , smail , i. , le borgne , j .- f . , blain , a. w. , kneib , j .-
p . , bezecourt , j. , kerr , t. h. , & davies , j. k. 1998 , , 298 , 583 ivison , r. j. , et al .
2002 , , 337 , 1 john , t. l. 1988 , , 193 , 189 kaspi , s. , smith , p. s. , netzer , h. , maoz , d. , jannuzi , b. t. , & giveon , u. 2000 , , 533 , 631 kawakatu , n. , umemura , m. , & mori , m. 2003 , , 583 , 85 kennicutt , r. c. 1998 , , 36 , 189 kormendy , j. , & richstone , d. 1995 , , 33 , 581 kuhlbrodt , b. , orndahl , e. , wisotzki , l. , & jahnke , k. , , submitted ledlow , m. j. , smail , i. , owen , f. n. , keel , w. c. , ivison , r. j. , & morrison , g. e. 2002 , , 577 , l79 leitherer , c. , et al . 1999 , , 123 , 3 magorrian , j. , et al .
1998 , , 115 , 2285 maraston , c. 2005 , , submitted marconi , a. , risaliti , g. , gilli , r. , hunt , l. k. , maiolino , r. , & salvati , m. 2004 , , 351 , 169 marconi , a. , & hunt , l. k. 2003 , , 589 , l21 martinez - sansigre , a. , rawlings , s. , lacy , m. , fadda , d. , marleau , f. r. , simpson , c. , willott , c. j. , & jarvis , m. j. 2005 , , in press mclure , r. j. , & dunlop , j. s. 2004 , , 352 , 1390 mihos , j. c. , & hernquist , l. 1994 , , 431 , l9 miller , g. e. , & scalo , j. m. 1979 , , 41 , 513 neri , r. , et al .
2003 , , 597 , l113 page , m. j. , stevens , j. a. , mittaz , j. p. d. , & carrera , f. j. 2001 , science , 294 , 2516 page , m. j. , stevens , j. a. , ivison , r. j. , & carrera , f. j. 2004 , , 611 , l85 poggianti , b. m. , & wu , h. 2000 , , 529 , 157 poggianti , b. m. , bressan , a. , & franceschini , a. 2001 , , 550 , 195 salpeter , e. e. 1955 , , 121 , 161 sanders , d. b. , soifer , b. t. , elias , j. h. , madore , b. f. , matthews , k. , neugebauer , g. , & scoville , n. z. 1988 , , 325 , 74 shapley , a. e. , steidel , c. c. , erb , d. k. , reddy , n. a. , adelberger , k. l. , pettini , m. , barmby , p. , & huang , j. , 2005 , , accepted shapley , a. e. , steidel , c. c. , adelberger , k. l. , dickinson , m. , giavalisco , m. , & pettini , m. 2001 , , 562 , 95 silk , j. , & rees , m. j. 1998 , , 331 , l1 simpson , c. , & eisenhardt , p. 1999
, , 111 , 691 smail , i. , chapman , s. c. , blain , a. w. , & ivison , r. j. 2004 , , 616 , 71 smail , i. , chapman , s. c. , ivison , r. j. , blain , a. w. , takata , t. , heckman , t. m. , dunlop , j. s. , & sekiguchi , k. 2003 , , 342 , 1185 smail , i. , ivison , r. j. , blain , a. w. , & kneib , j .-
2002 , , 331 , 495 smail , i. , ivison , r. j. , & blain , a. w. 1997 , , 490 , l5 steidel , c. c. , adelberger , k. l. , shapley , a. e. , pettini , m. , dickinson , m. , & giavalisco , m. 2003 , , 592 , 728 steidel , c. c. , shapley , a. e. , pettini , m. , adelberger , k. l. , erb , d. k. , reddy , n. a. , & hunt , m. p. 2004
, , 604 , 534 swinbank , a. m. , smail , i. , chapman , s. c. , blain , a. w. , ivison , r. j. , & keel , w. c. 2004 , , 617 , 64 swinbank , a. m. , et al .
2005 , , 359 , 401 van dokkum , p. g. , et al .
2003 , , 587 , l83 vernet , j. , & cimatti , a. 2001 , , 380 , 409 wandel , a. , peterson , b. m. , & malkan , m. a. 1999 , , 526 , 579 wang , w .- h . ,
cowie , l. l. , & barger , a. j. 2004 , , 613 , 655 webb , t. m. , et al .
2003 , , 587 , 41 worsley , m. a. , et al .
2005 , , 357 , 1281 | we analyze deep x - ray , optical and mid - infrared _ spitzer_observations of the cdf - n / goods
- n region to study a sample of 13 submillimeter - detected galaxies with spectroscopic redshifts ( median @xmath0 ) .
these galaxies are among the most active and massive at this epoch .
we find evidence for a power - law correlation between the estimated stellar and x - ray luminosity , implying that masses of the black holes may be related to the stellar masses of their host galaxies .
we derive the rest - frame uv near - infrared spectral energy distributions for these galaxies , believed to be young spheroids , and fit them with model templates . using the rest - frame near - infrared luminosities , which are relatively insensitive to uncertainties in stellar ages and reddening in these young dusty galaxies , and theoretical mass - to - light ratios
, we can estimate their stellar masses .
although the submillimeter emission implies that these galaxies are undergoing an epoch of intense star - formation , the _ spitzer _
data reveal a massive stellar population already in place .
we find that our submillimeter galaxies have a median stellar mass of @xmath1m@xmath2 , which is roughly ten times more massive than typical uv - selected star - forming systems at similar redshifts .
these stellar masses are then compared to previously published black hole mass estimates derived from the x - ray luminosities under the assumption of eddington - limit accretion .
we find that the black hole masses for our high - redshift sample are approximately 12 orders of magnitude smaller than galaxies of comparable stellar mass in the local universe .
although our estimates of black hole masses will increase if the accretion is sub - eddington , and our stellar masses will decrease if we assume a much younger stellar population or a different initial mass function , we find that only through a combination of effects is it possible to shift the high redshift galaxies such that they lie on the local relation .
this suggests that the black holes need to grow substantially between @xmath0 and the present - day , with much of the black hole growth occurring _ after _ the current obscured , far - infrared luminous phase of activity which is likely associated with the formation of the spheroid .
this interpretation supports a scenario where submillimeter galaxies pass through a subsequent accretion - dominated phase , where they would appear as optically bright quasars . |
the heavy dilepton pairs namely @xmath4 created in pb - pb collision at lhc energy is discussed in this work .
the major advantage of looking at @xmath4 dilepton pair arises due to the mass of the @xmath2 ( @xmath5 1.77 gev ) .
the @xmath2 pair mass distribution would then start beyond known contribution of hadronic resonances ( @xmath6 , @xmath7 and @xmath8 ) which dominates in the respective mass regions in @xmath9 and @xmath10 sector .
this would in turn mean the remaining contribution for @xmath2 production are due to thermal sources from partonic medium , pion annihilation in hadronic medium and drell yan mechanism .
we think that the results will definitely provide some useful baseline for experimental search and further detailed studies .
the main source of heavy dilepton @xmath11 pair production we have considered in this work is quark and anti - quark annihilation through photon , z and higgs bosons intermediated process .
the corresponding feynman diagrams are shown in fig .
they all contribute to thermal production of @xmath11 pair in quark gluon plasma .
the productions for these processes are evaluated from the matrix elements indicated below .
the matrix element for the process @xmath12 is given by , @xmath13[\bar{u(k_1)}{\gamma_\tau}v(k_2)]\ ] ] where , @xmath14 @xmath15[{c_v}^\tau-{c_a}^\tau{\gamma_5}]\ ] ] finally putting the values of the parameters we have , @xmath16 + s(0.858m_q^2 + 1.14m_\tau^2)\nonumber\\ & & -0.59m_q^2m_\tau^2 -0.0041\frac{s m_q^2 m_\tau^2}{m_z^2}+0.002 \frac{s^2 m_q^2 m_\tau^2}{m_z^2}\big ] \label{eq.z}\end{aligned}\ ] ] similarly , for the photon mediated process we have : @xmath17+(\frac{s}{2}-m_q^2)m_\tau^2\big ] \label{eq.photon}\ ] ] @xmath18 is the average charge of quarks , e is the electronic charge .
the matrix element of the interference term is @xmath19 .
@xmath20 + 0.0912\frac{s^3}{m_z^2 } \nonumber \\ & -&0.1824\frac{m_\tau^2}{m_z^2}s^2 + 0.5472\frac{m_q^2 m_\tau^2}{m_z^2}s + 0.1824 m_q^2 s-1.09 m_q^2 m_\tau^2\big ] \label{eq.interference}\end{aligned}\ ] ] finally , the matrix element for the higgs mediated process is : @xmath21[2m_\tau^2-\frac{s}{2 } ] \label{eq.h}\ ] ] here , @xmath22 , @xmath23 , @xmath24 , @xmath25 , are the masses of quarks , @xmath2 leptons , @xmath26 boson and higgs respectively .
( @xmath27 , @xmath28 ) and ( @xmath29 , @xmath30 ) are initial state and final state momenta respectively .
the total production cross section ( @xmath31 ) of @xmath32 is obtained by taking a coherent sum of the matrix elements given in eqs.([eq.z ] ) , ( [ eq.photon ] ) , ( [ eq.interference ] ) and ( [ eq.h ] ) with the following values of various parameters : @xmath24=91 gev , @xmath33=1.78 gev , @xmath25=120 gev , @xmath34=0.234 @xmath35=0.5 , @xmath36=0.19 , @xmath37=-0.5 , @xmath38=-0.03 and higgs vev=246 gev .
the total production cross - section @xmath31 is folded by the parton distribution functions to obtained the @xmath2 lepton pair yield in p - p collisions . in the present work cteq5
m pdf @xcite have been taken to obtain this .
the dy production of @xmath2 lepton in pb - pb collision is ; @xmath39 where , @xmath40 is the number of binary nucleon nucleon collisions an impact parameter b calculated using glauber model and @xmath41 is the inelastic cross section for p - p interaction .
we have taken @xmath42 mb and @xmath43 fm corresponding to @xmath44 centrality at @xmath45 tev .
the shadowing of parton distribution functions has been taken from @xcite .
the space time evolution of the system formed in pb+pb collisions at @xmath46 = 5.5 tev has been studied by using ideal relativistic hydrodynamics @xcite with longitudinal boost invariance @xcite and cylindrical symmetry .
our assummption is that the system reaches equilibration at a time @xmath47 = 0.08 fm / c after the collision . the initial temperature , @xmath48 is considered to be 700 mev and is calculated assuming the hadronic multiplicity ( dn / dy ) to be of the order of 2100 @xcite .
the equation of state ( eos ) obtained from the lattice qcd calculations by the milc collaboration @xcite for the partonic phase is used here . for the hadronic phase
eos all the resonances with mass 2.5 gev have been considered @xcite .
the transition temperature ( @xmath49 ) between hadronic phase and partonic phase is taken to be 175 mev @xcite .
we consider kinetic freeze out temperature , @xmath50 = 120 mev . in this context
we want to mention that the gluon fusion process was found to be dominant for mass of lepton pair greater than the mass of w boson .
our results are concentrated in the mass range of 4 to 20 gev , where the contribution from such process is found to be orders of magnitude smaller compared to the rest of the sources of @xmath11 pair production .
[ ! htb ] lepton pair yields as a function of invariant mass of the pair is displayed for pb+pb collisions at @xmath46 = 5.5 tev .
solid line indicates the spectra from quark gluon plasma and the dashed line stands for contribution from dy process . in ( b )
the ratio @xmath51 / @xmath52 $ ] is shown , here @xmath51 is the sum of the contribution shown in ( a ) , @xmath53 $ ] is the dy contribution from @xmath54 collision , and @xmath55 = 1369.477 for pb+pb collisions at @xmath46 = 5.5 tev.,title="fig : " ] lepton pair yields as a function of invariant mass of the pair is displayed for pb+pb collisions at @xmath46 = 5.5 tev .
solid line indicates the spectra from quark gluon plasma and the dashed line stands for contribution from dy process . in (
b ) the ratio @xmath51 / @xmath52 $ ] is shown , here @xmath51 is the sum of the contribution shown in ( a ) , @xmath53 $ ] is the dy contribution from @xmath54 collision , and @xmath55 = 1369.477 for pb+pb collisions at @xmath46 = 5.5 tev.,title="fig : " ] the yield ( @xmath56 ) for @xmath2 dilepton pair as a function of @xmath32 pair invariant mass for pb+pb collisions at @xmath46 = 5.5 tev is shown in figure [ fig2](a ) . the contributions from drell yan ( dy , dashed line ) and thermal partonic medium ( qgp , solid line ) are shown .
as evident from the figure , the drell yan contribution is higher than the thermal contribution for all the mass range studied .
the difference seems to increase with increase in @xmath32 pair mass .
figure [ fig2](b ) shows the ratio @xmath51 / @xmath52 $ ] , where @xmath51 is the sum of all the the contributions shown in figure [ fig2](a ) from pb+pb collisions .
the @xmath52 $ ] , is the number of binary collisions scaled contribution from dy process .
this contribution can be estimated from the measurement in p+p collisions at the same energy ( @xmath57 = 5.5 tev ) .
if there is no qgp formation then the ratio should always be equal to unity indicating the fact that the dilepton yield in the nucleus - nucleus collision is the collection of individual nucleon - nucleon collision only .
however , we observe that the ratio is above unity for the mass range of 4 to 6 gev . starting with a value of 1.4 at mass of 4 gev it decreases towards unity beyond mass of 6 gev .
this indicates that one should be able to extract a clear information of thermal contribution from partonic source at lhc energies using heavy dilepton pair measurement within the mass window of 4 to 6 gev .
a study of @xmath2 dilepton pair production at lhc energy has been carried out in this work . we have considered pb+pb collisions at midrapidity for @xmath46 = 5.5 tev .
it is expected that this energy should allow a significant production of @xmath2 leptons .
the main motivation of considering the @xmath2 lepton is its mass . because of heavy mass
the tau lepton yield is expected not to suffer from the huge background production .
the main sources for @xmath2 pair production considered here by quark and anti - quark annihilation mediated through photon , z and higgs boson .
the contribution from pion annihilation process is few orders of magnitude small compared to both thermal and drell yan contributions .
the drell yan contribution is found to be higher than the thermal contribution from partonic sources for the entire mass range studied .
the non - thermal contributions could be measured experimentally through p+p collisions , then the ratio of yields from nucleus - nucleus collisions to the yields for the binary collision scaled p+p collisions is found to be above unity for the mass range of 4 - 6 gev .
this indicates the window in mass region for @xmath2 dilepton pair where the thermal production can be studied at lhc energy using heavy dilepton pairs as an observable .
the support of prof .
jan e alam , vecc , dr .
b. mohanty , niser , prof .
sanjay k. ghosh and dr .
rajarshi ray of bose institute , kolkata is gratefully acknowledged . | we present a study of @xmath0 lepton pair production in pb + pb collisions at @xmath1 = 5.5 tev . the larger mass of tau lepton compared to electron and muon leads to considerably small hadronic contribution to the @xmath0 pair invariant mass ( m ) distribution relative to the production from thermal partonic sources .
the quark anti - quark annihilation processes via intermediary virtual photon , z and higgs bosons have been considered for the tau lepton production . the contribution from drell yan process
is found to dominate over thermal yield for @xmath0 pair mass from 4 to 20 gev at the lhc energy .
we also present the ratio of @xmath2 lepton pair yields for nucleus nucleus collisions relative to yields from p + p collisions scaled by number of binary collisions at lhc energies as a function @xmath2 lepton pair inavariant mass .
the ratio is found to be significantly above unity for the mass range 4 to 6 gev .
this indicates the possibility of detecting @xmath0 pair from quark
gluon plasma ( qgp ) in the mass window 4 to 6 gev .
* heavy dilepton in nucleus nucleus collision at lhc energy * + * sarbani majumder@xmath3 * + @xmath3department of physics , bose institute , kolkata , india |
given a digraph @xmath0 , the _ competition graph _
@xmath1 of @xmath0 is a graph which has the same vertex set as @xmath0 and has an edge between vertices @xmath2 and @xmath3 if and only if there exists a common out - neighbor of @xmath2 and @xmath3 in @xmath0 .
the notion of the competition graph is due to cohen @xcite and has arisen from ecology .
competition graphs also have applications in coding , radio transmission , and modeling of complex economic systems ( see @xcite ) .
since cohen introduced the notion of the competition graph , various variations have been defined and studied by many authors ( see @xcite and the survey articles @xcite ) .
cohen @xcite observed empirically that most competition graphs of acyclic digraphs representing food webs are interval graphs .
a graph @xmath4 is an _ interval graph _ if we can assign to each vertex @xmath3 in @xmath4 a real interval @xmath5 such that there is an edge between two distinct vertices @xmath3 and @xmath6 if and only if @xmath7 .
cohen s observation and the continued preponderance of examples that are interval graphs led to a large literature devoted to attempts to explain the observation and to study the properties of competition graphs .
roberts @xcite showed that every graph can be made into the competition graph of an acyclic digraph by adding isolated vertices .
he then asked for a characterization of acyclic digraphs whose competition graphs are interval .
the study of acyclic digraphs whose competition graphs are interval led to several new problems and applications ( see @xcite ) .
as one of the consequences , cho and kim @xcite found an interesting class of acyclic digraphs called doubly partial orders " with interval competition graphs .
we denote by @xmath8 the partial order @xmath9 on @xmath10 .
a digraph @xmath0 is called a _ doubly partial order _
( a dpo for short ) if there exist a finite subset @xmath11 of @xmath10 and a bijection @xmath12 such that @xmath13 .
the following theorem clarifies the relationship between interval graphs and the competition graphs of doubly partial orders .
[ thm;dpointkim ] the competition graph of a doubly partial order is an interval graph , and an interval graph with sufficiently many isolated vertices is the competition graph of a doubly partial order .
since then , it has been actively studied whether or not the same phenomenon occurs for other variants of competition graphs and interesting results have been obtained .
the competition - common enemy graph of a doubly partial order is an interval graph unless it contains a @xmath14-cycle as an induced subgraph .
in addition , an interval graph with sufficiently many isolated vertices is the competition - common enemy graph of a doubly partial order .
the above result on competition - common enemy graphs was generalized by lu and wu @xcite and wu and lu @xcite .
most recently , the niche graph , the @xmath15-step competition graph , and the phylogeny graph of a doubly partial order were studied .
the niche graph of a doubly partial order is an interval graph unless it contains a triangle . for any positive integer @xmath15 , the @xmath15-step competition graph of a doubly partial order
is an interval graph , and an interval graph with sufficiently many isolated vertices is the @xmath15-step competition graph of a doubly partial order .
the phylogeny graph of a doubly partial order is an interval graph .
in addition , for any interval graph @xmath4 , there exists an interval graph @xmath16 such that @xmath16 contains the graph @xmath4 as an induced subgraph and that @xmath16 is the phylogeny graph of a doubly partial order .
continuing in the same spirit , we study the competition hypergraph of a doubly partial order .
the notion of a competition hypergraph which is a variant of a competition graph was introduced by sonntag and teichert @xcite . the _ competition hypergraph _
@xmath17 of a digraph @xmath0 is a hypergraph without loops and multiple hyperedges such that the vertex set is the same as the vertex set of @xmath0 and @xmath18 is a hyperedge if and only if @xmath19 contains at least two vertices and @xmath19 coincides with the in - neighborhood of some vertex @xmath3 in @xmath0 .
as we study the competition hypergraphs of digraphs , we assume that all hypergraphs considered in this paper have no loops and no multiple hyperedges .
the notion of a competition hypergraph is considered as one of the important variants of competition graphs and significant results on this topic are being obtained ( see @xcite ) . in this paper , we classify doubly partial orders whose competition hypergraphs are interval .
throughout this section , we follow the terminology for hypergraphs given in @xcite .
we say that two vertices @xmath2 and @xmath3 are _ adjacent _ in a hypergraph @xmath20 if there is a hyperedge @xmath19 in @xmath20 such that @xmath21 . for a positive integer @xmath22 , a hypergraph @xmath20 is called _ @xmath22-uniform _ if each hyperedge of the hypergraph @xmath20 has the same size @xmath22 .
obviously , @xmath23-uniform hypergraphs are graphs .
a sequence @xmath24 of distinct vertices of a hypergraph @xmath20 is called a _ path _
if there exist @xmath25 distinct hyperedges @xmath26 such that @xmath27 contains @xmath28 for each @xmath29 .
a sequence @xmath24 of distinct vertices of a hypergraph @xmath20 is called a _ cycle _
if there exist @xmath30 distinct hyperedges @xmath31 such that @xmath27 contains @xmath28 for each @xmath32 and @xmath33 contains @xmath34 .
a _ subhypergraph _ of a hypergraph @xmath20 is a hypergraph @xmath35 such that @xmath36 and @xmath37 . for a vertex @xmath3 in a digraph @xmath0 ,
we denote by @xmath38 the in - neighborhood of @xmath3 , i.e. , @xmath39 .
a hypergraph @xmath20 is _ interval _ if there exists a one - to - one function mapping the vertices of @xmath40 to points on the real line such that for each hyperedge @xmath19 , there exists an interval containing the images of all elements of @xmath19 , but not the images of any vertices not in @xmath19 .
there is a characterization of interval hypergraphs by forbidden subhypergraphs : [ thm : forbidden ] a hypergraph @xmath20 is an interval hypergraph if and only if @xmath20 does not contain any of the hypergraphs in figure [ fig1 ] as a subhypergraph .
more precisely , the hypergraphs in figure [ fig1 ] are defined as follows : given a positive integer @xmath41 , let @xmath42 be the @xmath23-uniform hypergraph with @xmath43 vertices which forms a cycle , and let @xmath44 .
for a positive integer @xmath43 , we define hypergraphs @xmath45 and @xmath46 with @xmath47 vertices by @xmath48 let @xmath49 and @xmath50 .
let @xmath51 be the hypergraph defined by @xmath52 and @xmath53 , and let @xmath54 be the hypergraph defined by @xmath55 and @xmath56 .
theorem [ thm : forbidden ] states that a hypergraph @xmath20 being an interval hypergraph is equivalent to @xmath20 not containing any of the hypergraphs in @xmath57 as a subhypergraph .
first , we will show that the competition hypergraph of a dpo may not be interval .
we will always embed the vertices of a dpo ( as well as the vertices of its competition hypergraph ) into @xmath10 in a natural way . for a positive integer @xmath43
, we define @xmath58 in the dpo defined on the set @xmath59 , two vertices @xmath60 and @xmath61 of @xmath62 with @xmath63 have a common out - neighbor @xmath64 if @xmath65 and have no common out - neighbor if @xmath66 .
thus , the competition hypergraph of the dpo defined on the set @xmath59 is a path as a @xmath23-uniform hypergraph on the @xmath67 vertices in @xmath62 together with the @xmath68 isolated vertices in @xmath69 .
[ lem;m ] for a positive integer @xmath43 , there exists a doubly partial order whose competition hypergraph contains @xmath45 as a subhypergraph . for a positive integer @xmath43 , we will define a dpo @xmath70 such that @xmath71 contains @xmath45 as a subhypergraph .
it is easy to check that , for the dpo @xmath72 defined on the set @xmath73 ( see figure [ fig2 ] ) , @xmath74 contains @xmath75 as a subhypergraph . for a positive integer @xmath76 , let @xmath70 be the dpo defined by @xmath77 ( see figure [ fig2_1 ] ) .
then the hyperedges of @xmath71 consist of the hyperedges of the @xmath23-uniform path induced by @xmath62 and the hyperedge @xmath78 .
note that @xmath79 for @xmath76 .
thus , it is easy to see that the subhypergraph of @xmath71 induced by @xmath80 is isomorphic to @xmath45 .
[ lem;f ] for a positive integer @xmath43 , there exists a doubly partial order whose competition hypergraph contains @xmath46 as a subhypergraph . for a positive integer @xmath43 , we will define a dpo @xmath81 such that @xmath82 contains @xmath46 as a subhypergraph .
it is easy to check that , for the dpo @xmath83 defined on the set @xmath84 ( see figure [ fig4 ] ) , @xmath85 contains @xmath86 as a subhypergraph . for @xmath76 ,
let @xmath81 be the dpo @xmath81 defined by @xmath87 ( see figure [ fig4_1 ] ) .
then the hyperedges of @xmath82 consist of the hyperedges of the 2-uniform path induced by @xmath62 and the hyperedges @xmath88 and @xmath89 .
note that @xmath90 for @xmath76 .
thus , it is easy to see that the subhypergraph of @xmath82 induced by @xmath91 is isomorphic to @xmath46 .
lemmas [ lem;m ] and [ lem;f ] show that a hypergraph isomorphic to an element in @xmath92 is realizable as a subhypergraph of the competition hypergraph of a dpo .
[ main;interval ] for each hypergraph @xmath20 in @xmath92 , there exists a doubly partial order whose competition hypergraph contains @xmath20 as a subhypergraph . now it is natural to ask whether the family @xmath92 contains all the forbidden subhypergraphs for the competition hypergraph of a dpo being interval .
the answer is yes , as it will be shown in the remainder of this paper . for the sake of simplicity
, we define an irreflexive and transitive relation @xmath93 on @xmath10 as follows : for @xmath94 with @xmath95 and @xmath96 , @xmath97 obviously , for @xmath98 , @xmath99 with @xmath100 , @xmath101 the following lemmas are simple but useful : [ lem;prec ] for vertices @xmath98 and @xmath102 of a doubly partial order @xmath0 , if the competition hypergraph of @xmath0 has two hyperedges @xmath103 and @xmath104 such that @xmath105 , @xmath106 , @xmath107 , @xmath108 , then either @xmath109 or @xmath110 . to reach a contradiction ,
suppose that @xmath111 and @xmath112 .
then , by ( @xmath113 ) , either @xmath114 or @xmath115 . if @xmath114 , then any out - neighbor of @xmath98 is also an out - neighbor of @xmath102 .
therefore , any hyperedge containing @xmath98 contains @xmath102 , which is a contradiction to the existence of the hyperedge @xmath103 .
similarly , we can reach a contradiction if @xmath115 .
thus @xmath109 or @xmath110 .
[ lem;prec2 ] for vertices @xmath98 , @xmath102 and @xmath116 of a doubly partial order @xmath0 , if @xmath117 then @xmath102 and all of its in - neighbors are contained in any hyperedge containing @xmath98 and @xmath116 in the competition hypergraph of @xmath0 .
let @xmath19 be a hyperedge containing @xmath98 and @xmath116 in @xmath17 .
then there exists a vertex @xmath118 such that @xmath119 .
let @xmath95 , @xmath96 , @xmath120 , and @xmath121 . since @xmath122 and @xmath123 , @xmath124 and @xmath125 .
since @xmath117 , @xmath126 and @xmath127 , which implies that @xmath128 and @xmath129 .
therefore @xmath130 and thus @xmath131 .
since @xmath132 for any in - neighbor @xmath2 of @xmath102 , @xmath133 and hence @xmath134 . [ lem;prec3 ] for vertices @xmath98 , @xmath102 , @xmath116 of a doubly partial order @xmath0 , if @xmath135 , @xmath136 , and @xmath137 , then @xmath138 .
let @xmath95 , @xmath96 , and @xmath120 .
since @xmath135 , @xmath139 . since @xmath136 , @xmath140 .
therefore @xmath141 and so @xmath142 .
since @xmath142 and @xmath143 , either @xmath138 or @xmath144 by ( @xmath113 ) . if @xmath144 , then @xmath145 and so @xmath146 , a contradiction to the fact that @xmath136 .
thus @xmath147 and so @xmath138 .
[ lem;prec4 ] for vertices @xmath98 , @xmath102 , @xmath116 of a doubly partial order @xmath0 , if @xmath135 , @xmath148 , and @xmath149 , then @xmath150 .
let @xmath95 , @xmath96 , and @xmath120 .
since @xmath135 , @xmath151 . since @xmath148 , @xmath152 . therefore @xmath153 . since @xmath149 , @xmath154 or @xmath155 . since @xmath153 , @xmath154 . thus @xmath150
. using lemmas [ lem;prec ] , [ lem;prec2 ] , and [ lem;prec3 ] , we can show that the competition hypergraph of a dpo does not contain any element of @xmath156 as a subhypergraph .
[ lem;c ] the competition hypergraph of a doubly partial order does not contain any hypergraph in @xmath157 as a subhypergraph .
we prove the lemma by contradiction .
suppose that there exists a dpo @xmath0 such that @xmath17 contains @xmath42 as a subhypergraph for some @xmath158 .
let @xmath159 be the vertices of @xmath42 such that @xmath160 , @xmath161 are adjacent for all @xmath162 , where the subscripts are reduced modulo @xmath43 .
note that for any distinct @xmath163 , @xmath164 in @xmath165 , there exists a hyperedge containing @xmath166 but not @xmath167 .
thus , by lemma [ lem;prec ] , @xmath168 or @xmath169 for any distinct @xmath163 , @xmath164 in @xmath165 . without loss of generality
, we may assume that @xmath170 .
if @xmath171 , then @xmath172 and so , by lemma [ lem;prec2 ] , @xmath173 and @xmath174 are adjacent , a contradiction .
thus @xmath175 .
suppose that @xmath176 .
then @xmath177 , and by the argument in the proof of lemma [ lem;prec2 ] , a common out - neighbor of @xmath173 and @xmath178 is also an out - neighbor of @xmath174 .
this implies that @xmath42 contains a hyperedge containing @xmath179 , which contradicts that @xmath42 is @xmath23-uniform .
therefore , @xmath180 . by applying a similar argument
, we can claim that @xmath181 . by continuing this argument
, we obtain @xmath182 .
since the relation @xmath93 is transitive , we have @xmath183 , which is a contradiction to the irreflexivity of @xmath93 .
[ lem;o ] the competition hypergraph of a doubly partial order does not contain @xmath51 or @xmath54 as a subhypergraph .
suppose that there exists a dpo @xmath0 such that @xmath17 contains @xmath51 as a subhypergraph .
let @xmath184 be the hyperedge of @xmath51 of size @xmath185 .
by lemma [ lem;prec ] , any two @xmath2 , @xmath3 of @xmath186 satisfy @xmath187 or @xmath188 .
thus @xmath189 for a permutation @xmath190 on @xmath186 .
without loss of generality , we may assume that @xmath117 .
let @xmath191 be the vertex of @xmath51 other than @xmath192 that is adjacent to @xmath102 .
by lemma [ lem;prec ] , either @xmath193 or @xmath194 , and either @xmath195 or @xmath196 .
if @xmath194 , then @xmath197 and so , by lemma [ lem;prec2 ] , @xmath98 and @xmath191 are adjacent , a contradiction . thus @xmath193 .
if @xmath196 , then @xmath198 and so , by lemma [ lem;prec2 ] , @xmath98 and @xmath191 are adjacent , a contradiction .
thus @xmath195 . then @xmath199 and so , by lemma [ lem;prec2 ] , @xmath191 and @xmath116 are adjacent , a contradiction
hence , the competition hypergraph of a dpo does not contain @xmath51 as a subhypergraph .
suppose that there exists a dpo @xmath0 such that @xmath17 contains @xmath54 as a subhypergraph .
let @xmath200 and let @xmath201 be the hyperedge of @xmath54 of size @xmath14 , and @xmath202 and @xmath203 be the hyperedges of size @xmath23 .
the vertices @xmath102 and @xmath116 satisfy the condition of lemma [ lem;prec ] and so @xmath144 or @xmath138 . without loss of generality
, we may assume that @xmath144 .
now @xmath98 and @xmath116 satisfy the condition of lemma [ lem;prec ] and so @xmath150 or @xmath204 . if @xmath204 then @xmath205 and so @xmath116 belongs to any hyperedge containing @xmath98 and @xmath102 by lemma [ lem;prec2 ] , a contradiction . therefore @xmath150 .
similarly we can show that @xmath206 . since @xmath207
, there exist @xmath118 , @xmath208 such that @xmath209 and @xmath210 .
if @xmath211 , then @xmath212 , a contradiction .
if @xmath213 , then @xmath214 , a contradiction .
therefore @xmath215 and @xmath216 .
thus @xmath102 , @xmath6 , @xmath118 satisfy the condition of lemma [ lem;prec3 ] and so @xmath217 . if @xmath218 then @xmath219 , a contradiction . therefore @xmath220 , and so @xmath98 , @xmath116 , @xmath118 satisfy the condition of lemma [ lem;prec4 ] and so @xmath221 . thus @xmath222 . since @xmath223 is a hyperedge containing @xmath98 and @xmath6 , the vertex @xmath3 which is an in - neighbor of @xmath118 belongs to @xmath223 by lemma [ lem;prec2 ]
since @xmath224 , @xmath225 .
however , @xmath226 , a contradiction .
hence the competition hypergraph of a dpo does not contain @xmath54 as a subhypergraph .
by lemmas [ lem;c ] and [ lem;o ] , the following holds : [ thm;main ] for a doubly partial order @xmath0 , the competition hypergraph of @xmath0 is interval if and only if it does not contain any hypergraph in @xmath227 as a subhypergraph .
it follows from theorems [ main;interval ] and [ thm;main ] that @xmath92 is the family of all the forbidden subhypergraphs for the competition hypergraph of a dpo being interval .
a hypergraph is _ chordal _ if any cycle of length at least @xmath14 has two nonconsecutive vertices that are adjacent . if a hypergraph does not contain any hypergraph in @xmath157 as a subhypergraph , then it is chordal
. therefore , by lemma [ lem;c ] , the following corollary immediately holds .
the competition hypergraph of a doubly partial order is chordal .
theorem [ thm;dpointkim ] shows that any interval graph can be made into the competition graph of a dpo by adding sufficiently many isolated vertices . in the same context
, we may ask whether the following statement is true : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * an interval hypergraph can be made into the competition hypergraph of a dpo by adding sufficiently many isolated vertices . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ we can show that the answer is no by taking an interval hypergraph @xmath20 defined by @xmath228 and @xmath229 ( see figure [ fig : inthyp ] ) . to show by contradiction that @xmath20 can not be made into the competition hypergraph of a dpo even if sufficiently many isolated vertices are added , suppose that there is a dpo @xmath0 such that @xmath230 where @xmath231 is a set of isolated vertices .
one can observe that any two distinct vertices among @xmath178 , @xmath174 , @xmath232 , @xmath233 satisfy the condition in lemma [ lem;prec ] .
thus , for any two distinct vertices @xmath98 , @xmath102 in the hyperedge @xmath234 , either @xmath109 or @xmath110 holds . by applying a similar argument for assuming that @xmath235 in the proof of lemma [ lem;c ]
, we may assume without loss of generality that @xmath236 .
since @xmath237 , there exists a vertex @xmath2 such that @xmath238 . if @xmath239 for @xmath240 , then @xmath174 and @xmath232 are contained in any hyperedge containing @xmath160 , a contradiction .
thus @xmath241 .
since @xmath180 , @xmath242 , and @xmath243 , it follows from lemma [ lem;prec4 ] that @xmath244 . on the other hand , since @xmath245 , @xmath246 , and @xmath247 , it follows from lemma [ lem;prec3 ] that @xmath248 .
we have shown that @xmath249 .
then , by lemma [ lem;prec2 ] , @xmath2 is contained in the hyperedge @xmath250 , which is a contradiction to @xmath241 . the following statement weaker than @xmath251 seems to be worthy of mention : [ thm;main2 ] for an interval hypergraph @xmath20
, there exists a doubly partial order @xmath0 whose competition hypergraph contains @xmath20 as a subhypergraph .
let @xmath20 be an interval hypergraph .
then there exists an ordering @xmath252 of the vertices of @xmath20 such that any hyperedge of @xmath20 is consecutive in the ordering . for each @xmath253 , let @xmath254 . for each hyperedge
@xmath19 , we define @xmath255 and @xmath256 and let @xmath257 .
let @xmath0 be a dpo on the set @xmath258 , and let @xmath35 be the subhypergraph of @xmath17 induced by the set @xmath259 .
we will show that the bijection from @xmath40 to @xmath260 which maps a vertex @xmath160 to the vertex @xmath261 is an isomorphism between hypergraphs @xmath20 and @xmath35 .
let @xmath19 be a hyperedge in @xmath20 , and let @xmath262 and @xmath263 . then @xmath264 .
since @xmath19 contains at least two vertices , @xmath265 . then @xmath266 is a vertex of @xmath0 . if @xmath267 , then @xmath268 and @xmath269 . if @xmath270 , then @xmath271 . if @xmath272 , then @xmath273
therefore , we have @xmath274 if @xmath267 and @xmath275 if @xmath270 or @xmath272 .
since @xmath276 is a hyperedge of @xmath17 , @xmath277 is a hyperedge in @xmath35 .
let @xmath278 be a hyperedge in @xmath35 .
let @xmath279 and @xmath280 .
note that @xmath281 for @xmath270 and @xmath272 .
since @xmath278 contains at least two vertices , @xmath282 . then there exists a vertex @xmath283 such that @xmath284 .
note that @xmath285 and @xmath286 . if @xmath267 , then @xmath287 since @xmath288 and @xmath289 .
therefore , @xmath290 . by the definition of @xmath0 and the fact that @xmath291 for all @xmath292 , @xmath293 for some hyperedge @xmath19 of @xmath20 . by the consecutive property of the ordering @xmath252
, the hyperedge @xmath19 contains @xmath167 for each @xmath164 such that @xmath294 .
now we show that @xmath295 and @xmath296 .
since @xmath297 and @xmath298 , we obtain @xmath299 and @xmath300 .
since @xmath301 , it holds that @xmath302 or @xmath303 .
if the latter happens , then @xmath304 , which contradicts the choice of @xmath163 .
thus @xmath305 .
since @xmath306 , it holds that @xmath307 or @xmath308 . by the choice of @xmath25 , the latter holds and so @xmath309 .
thus @xmath295 and @xmath296 .
hence the theorem holds .
the first author s work was supported by a national research foundation of korea ( nrf ) grant funded by the korean government ( mest ) ( no .
2011 - 0005188 ) .
the third author s research was supported by a national research foundation of korea grant funded by the korean government , the ministry of education , science and technology ( nrf-2011 - 357-c00004 ) .
the fourth author was supported by jsps research fellowships for young scientists .
k. f. fraughnaugh , j. r. lundgren , j. s. maybee , s. k. merz , n. j. pullman : competition graphs of strongly connected and hamiltonian digraphs , _ siam journal on discrete mathematics _ * 8 * ( 1995 ) 179185 .
j. r. lundgren : food webs , competition graphs , competition - common enemy graphs , and niche graphs , _ applications of combinatorics and graph theory to the biological and social sciences _ ( the i m a volumes in mathematics and its applications * 17 * , springer - verlag , new york , 1989 ) 221243 . a. raychaudhuri , f. s. roberts : generalized competition graphs and their applications , _ ix symposium on operations research . part i. sections 14 ( osnabrck , 1984 ) _ ( methods of operations research * 49 * , athenum / hain / hanstein , knigstein , 1985 ) 295311 .
f. s. roberts : food webs , competition graphs , and the boxicity of ecological phase space , _ theory and applications of graphs ( proc .
, western mich .
kalamazoo , mich . , 1976 ) _
( lecture notes in mathematics * 642 * , springer , new york , 1978 ) 477490 .
f. s. roberts : competition graphs and phylogeny graphs , in l. lovasz ( ed . ) , _ graph theory and combinatorial biology ( balatonlelle , 1996 ) _
( bolyai society mathematical studies * 7 * , jnos bolyai math .
budapest , 1999 ) 333362 . | since cho and kim ( 2005 ) @xcite showed that the competition graph of a doubly partial order is an interval graph , it has been actively studied whether or not the same phenomenon occurs for other variants of competition graphs and interesting results have been obtained . continuing in the same spirit ,
we study the competition hypergraph , an interesting variant of the competition graph , of a doubly partial order .
though it turns out that the competition hypergraph of a doubly partial order is not always interval , we completely characterize the competition hypergraphs of doubly partial orders which are interval .
competition hypergraphs , competition graphs , doubly partial orders , interval hypergraphs 05c75 , 05c20 |
according to the recent cosmic observations our universe is experiencing an accelerating phase of expansion at the present time @xcite-@xcite . since the ordinary barionic matter ( energy ) can only produces an attractive force , there should necessarily an exotic form of energy with negative pressure called dark energy " ( de ) be exist in order to drive the observed cosmic acceleration @xcite-@xcite .
moreover , based on the recent observations we live in a nearly spatially flat universe composed of approximately @xmath8 baryonic matter , @xmath9 dark matter and @xmath10 dark energy .
a natural candidate for dark energy is cosmological constant @xmath11 , which has the equation of state @xmath12 .
however , although cosmological constant can explain the present cosmic acceleration , it would encounter some serious theoretical problems , such as the fine - tuning and the coincidence problems .
another possible candidates for dark energy are the dynamic dark energy models provided by scalar fields such as quintessence @xmath13 @xcite-@xcite , phantom ( @xmath14 ) @xcite , quintom ( @xmath15 ) @xcite , k - essence @xcite , chaplygin gas as well as generalized chaplygin gas models @xcite , and etc .
+ we also can study dark energy through an almost model - independent way . in this approach
we parameterize the dark energy eos parameter by giving the concrete form of the equation of state of dark energy directly , such as @xmath16 @xcite , @xmath17 @xcite , etc .
studies of dark energy as a function of redshift , by using this method , show that though the cosmological constant is not ruled out in 1@xmath18 region , the current constraints favors a dynamical dark energy .
however , as noted in ref @xcite the rapid changed of eos parameter i.e @xmath19 is ruled out .
moreover , it is a well established fact that the universe is accelerated expansion at present ( dark energy dominated ) and decelerated expansion in the past ( dark matter dominated ) .
therefore , the deceleration parameter ( dp ) should not be a constant but time ( or redshift ) variable which is positive in the past and changes into negative at @xmath20 @xcite-@xcite .
considering the idea of variable dp , the parameterized decelerated parameter is present in almost model independent way by giving a concrete form of decelerated parameters .
+ the high symmetry involved in flrw models requires a very high degree of fine tuning of initial conditions which is extraordinary improbable and hence flrw models are infinitely improbable in the space of all possible cosmologies @xcite .
although there are an increasing interest in the study of de in the scop of anisotropic space - times , there is not a reference paper including the present observational constraints on the observable parameters such as @xmath1 , @xmath21 , @xmath4 , etc .
therefore , still authors compare their results by those obtained on the bases of the standard cosmological models .
motivated the situation discussed above , in this letter instead of flrw metric , an anisotropic space - time namely bianchi type i metric is used in order to obtain more general results .
up to our knowledge , such kind of calculations has not been done yet .
first we derive the general form of the eos parameter in bianchi type i space - time and then we use the hubble dataset , cmb data , and their combination to put the observational constraints on the model parameters .
i is worth to mention that because of technical difficulties in minimizing @xmath22 the results obtained in this work are not so general and more work is needed to extend these results for example by minimizing @xmath23 .
we consider the bianchi type i space - time in the orthogonal form as @xmath24 where @xmath25 and @xmath26 are functions of time only .
+ the einstein s field equations ( in gravitational units @xmath27 ) read as @xmath28 where @xmath29 and @xmath30 are the energy momentum tensors of barotropic matter and dark energy , respectively .
these are given by @xmath31,\ ] ] @xmath32\rho^{m},\ ] ] and @xmath33,\ ] ] @xmath34\rho^{x},\ ] ] where @xmath35 and @xmath36 are , respectively the energy density and pressure of the perfect fluid component or ordinary baryonic matter while @xmath37 is its eos parameter .
similarly , @xmath38 and @xmath39 are , respectively the energy density and pressure of the de component while @xmath40 is the corresponding eos parameter .
we assume the four velocity vector @xmath41 satisfying @xmath42 . + in a co - moving coordinate system ( @xmath43 ) , einstein s field equations ( [ eq2 ] ) with ( [ eq3 ] ) and ( [ eq4 ] ) for b - i metric ( [ eq1 ] ) subsequently lead to the following system of equations : @xmath44 @xmath45 @xmath46 @xmath47 if we consider @xmath48 as the average scale factor of bianchi type i model , then the generalized mean hubble s parameter @xmath49 defines as @xmath50 the bianchi identity @xmath51 leads to @xmath52 .
therefore , the continuity equation for dark energy and baryonic matter can be written as @xmath53 solving eqs .
( [ eq5]-([eq8 ] ) one can find @xcite @xmath54 @xmath55 and @xmath56 where @xmath57 using eqs .
( [ eq11])-([eq13 ] ) in eqs .
( [ eq5])-([eq8 ] ) we can write the analogue of the friedmann equation as @xmath58 and @xmath59 here @xmath60 , @xmath61 and @xmath62 .
note that @xmath63 denotes the deviation from isotropy e.g. @xmath64 represents flat flrw universe .
thus , when the universe is sufficiently large , almost at the present time , the space - time ( [ eq1 ] ) behaves like a flat flrw universe .
+ moreover , if there is no interaction between dark energy and cold dark matter(cdm ) with @xmath65 , then one can write the conservation equation ( [ eq10 ] ) for cdm and dark energy separately as separately as @xmath66 and @xmath67 eq.([eq16 ] ) leads to @xmath68 using eqs .
( [ eq14 ] ) , ( [ eq18 ] ) in eqs .
( [ eq7 ] ) and ( [ eq8 ] ) we obtain the energy density and pressure of dark fluid as @xmath69 and @xmath70 respectively . here
, @xmath71 is a positive constant ( note that @xmath72 ) .
therefore , the equation of state parameter ( eos ) of de in it s general form is given by @xmath73 where @xmath74 is the deceleration parameter and @xmath75 is the current value of matter density .
in this section , by using @xmath22 method we investigate the constraints on the parameters @xmath0 , @xmath76 , and @xmath4 .
to do so we utilize the recent observational data including hubble parameter and the cosmic microwave background ( cmb ) radiation .
note that as mentioned by kumatsu et al @xcite the baryonic oscillation ( bao ) distance ratio is not applicable in non - flrw based models .
+ first of all , from eq .
( [ eq21 ] ) we obtain the expression for the hubble rate , of the form @xmath77 in our study , following xu l x et al .
@xcite , the deceleration parameter is considered to be @xmath78 where @xmath79 and @xmath80 are constants determined by the recent observational constraints . from this equation
it is obvious that for @xmath81 , @xmath82 which is corresponding to matter dominated era whereas for @xmath83 , the current value of deceleration parameter is obtained as @xmath84 . in view of eq .
( [ eq23 ] ) , the hubble parameter is written in the form @xmath85,\ ] ] we also consider the following simple expression for the equation of state @xmath86 where @xmath87 and @xmath88 are constants .
using eqs .
( [ eq23 ] ) and ( [ eq25 ] ) one can easily re - write eq .
( [ eq22 ] ) as a function of @xmath89 only .
+ as far as we are dealing with the experimental @xmath90 test , it has been suggested by jimenez et .
al . @xcite to use the quantity @xmath91 which is called _
differential age_. in this method , one can directly analyze the @xmath90 data without passing through the luminosity distance as we usually do in the case of supernovae ia . using the differential ages of passively - evolving galaxies from the gemini deep deep survey ( gdds)@xcite and archival data @xcite ,
first , a set of nine values of @xmath90 in the range of @xmath92 has been reported @xcite .
table 1 shows the most recent data set includes 14 values .
+ we minimize the following reduced @xmath22 @xcite to constrain the parameters @xmath0 and @xmath76 for fixed @xmath93 at @xmath94 and @xmath95 .
@xmath96^{2}}{\sigma^{2}_{hub}(z_{i})},\ ] ] where @xmath97 are the values of table 1 and @xmath98 .
+ .the cosmological data at 1@xmath18 error for @xmath90 expressed in @xmath99 . [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] [ table : nonlin ] the @xmath100 contours of parameters @xmath79 and @xmath80 , the variation of the decelaration parameter @xmath1 , and the @xmath100 contours of parameters @xmath88 and @xmath87 are depicted in figures @xmath101 , and @xmath102 respectively .
all these figures show that the combined constraint is stricter than independent constraints given by cmb dataset or the hubble data .
+ 0.3 and @xmath80 with @xmath100 confidence level according to : ( a ) hubble , ( b ) cmb , and ( c ) h + cmb combination data , title="fig : " ] 0.3 and @xmath80 with @xmath100 confidence level according to : ( a ) hubble , ( b ) cmb , and ( c ) h + cmb combination data , title="fig : " ] 0.3 and @xmath80 with @xmath100 confidence level according to : ( a ) hubble , ( b ) cmb , and ( c ) h + cmb combination data , title="fig : " ] 0.3 versus red shift z , which are constrained according to : ( a ) hubble , ( b ) cmb , and ( c ) h + cmb combination data , title="fig : " ] 0.3 versus red shift z , which are constrained according to : ( a ) hubble , ( b ) cmb , and ( c ) h + cmb combination data , title="fig : " ] 0.3 versus red shift z , which are constrained according to : ( a ) hubble , ( b ) cmb , and ( c ) h + cmb combination data , title="fig : " ] 0.3 obtained by : ( a ) hubble , ( b ) cmb , and ( c ) h + cmb combination data .
a cross sing in figures shows cosmological constant corresponds to @xmath103.,title="fig : " ] 0.3 obtained by : ( a ) hubble , ( b ) cmb , and ( c ) h + cmb combination data .
a cross sing in figures shows cosmological constant corresponds to @xmath103.,title="fig : " ] 0.3 obtained by : ( a ) hubble , ( b ) cmb , and ( c ) h + cmb combination data .
a cross sing in figures shows cosmological constant corresponds to @xmath103.,title="fig : " ] the best fit transition redshift @xmath4 for the case of combined constraint at the @xmath100 error is @xmath7 .
it is worth to mention that the constraints to the deceleration parameter @xmath1 and the dark energy equation of state are looser in 1@xmath18 error by using hubble data than cmb data , which may be attributed to a fewer data points from hubble data .
+ the variations of the dp parameter @xmath1 and the eos parameter @xmath0 are plotted in fig .
we have used the best fit values of @xmath79 , @xmath80 , @xmath87 , and @xmath88 from table 2 . and @xmath104 versus redshift ( @xmath89 ) for @xmath105 , and @xmath106 .
these values are taken from the combined data in table 2.,width=377,height=377 ]
this work has been supported by a research fund from the mahshahr branch of islamic azad university under the project entitled the role of scalar fields in the study of dark energy " .
author would like to thank s. darvishi for his valuable help in calculating some parameters of table 2 .
000 s perlmutter , _ et al , astrophys .
j. _ * 517 * 565 ( 1999 ) a g riess _ et al , astron . j. _ * 116 * , 1009 ( 1998 ) a g riess _ et al , astrophys . j. _ * 560 * ,
49 ( 2001 ) j l tonry _ et al , astrophys .
j. _ * 594 * , 1 ( 2003 ) m tegmark _ et al , phys .
* d69 * , 103501 ( 2004 ) c l bennet _
et al , astrophys . j. suppl . _ * 148 * , 1 ( 2003 ) d n spergel _
et al , astrophys . j. suppl . _
* 148 * , 175 ( 2003 ) k abazajian
_ et al , astron .
j. _ * 128 * , 502 ( 2004 ) b ratra and p. j. e. peebles , _ phys .
rev . _ * d37 * , 3406 ( 1988 ) c wetterich , _ nucl .
phys . _ * b302 * , 668 ( 1988 ) m s turner and m. j. white , _ phys
. rev . _ * d56 * , 4439 ( 1997 ) r r caldwell , r. dave and p. j. steinhardt , _ phys .
lett . _ * 80 * , 1582 ( 1998 ) a r liddle and r. j. scherrer , _ phys . rev . _ * d59 * , 023509 ( 1998 )
p j steinhardt , l. m. wang and i. zlatev , _ phys .
rev . _ * d59 * , 123504 ( 1999 ) r r caldwell , _ phys
_ * b545 * , 23 ( 2002 ) b feng , x l wang , and x zhang , _ phys .
lett . _ * b607 * , 35 ( 2005 ) a picon , t damour and v mukhanov , _ phys .
* b558 * , 209 ( 1999 ) m malquarti , e j copeland , a r liddle and m trodden , _ phys .
rev . _ * d67 * , 123503 ( 2003 ) s k srivastava , _ phys .
lett . _ * b619 * , 1 ( 2005 ) u alam , v sahni , t d saini and a a starobinsky , _ mon . not .
soc . _ * 344 * , 1057 ( 2003 ) a r cooray and d. huterer , _ astrophys . j. _ * 513 * , l95 ( 1999 ) m chevallier , d polarski , _ int
_ * d10 * , 213 ( 2001 ) e v linder , _ phys .
lett . _ * 90 * , 091301 ( 2003 ) a g riess , _ et al , astrophys . j. _ * 659 * , 98 ( 2007 ) l xu , h liu and y ping , _ int .
* 45 * , 869(2006 ) n banerjee , s das , _ gen.rel.grav . _ * 37 * , 1695 ( 2005 ) m goliath and g f r ellis , _ phys
* d60 * , 032502 ( 1999 ) b saha , _ mod .
lett . _ * a20 * , 2127 ( 2005 ) e komatsu , _ et al , astrophys . j. suppl . _ * 192 * , 18 ( 2011 ) l xu , h liu and y ping , _ int . j. theor
* 45 * , 843 ( 2006 ) r jimenez , l verde , t treu and d stern , _ astrophys .
j. _ * 593 * , 622 ( 2003 ) r g abraham , _
et al , astron . j. _ * 127 * , 2455 ( 2004 ) t treu , _ et al , mon . not .
soc . _ * 308*,1037 ( 1999 ) t treu , _ et al , mon . not .
soc . _ * 326 * , 221 ( 2001 ) simon j , _ et al , phys .
_ * d71 * , 123001 ( 2005 ) o luongo , _ mod .
* a26 * , 1459 ( 2011 ) | we study the constraints on dark energy equation of state @xmath0 and the deceleration parameter @xmath1 from the recent observational data including hubble data and the cosmic microwave background ( cmb ) radiation by using a model - independent deceleration parameter @xmath2 and dark energy equation of state @xmath3 in the scope of anisotropic bianchi type i space - time . for the cases of hubble dataset , cmb data , and their combination , our results indicate that the constraints on transition redshift @xmath4 are @xmath5 , @xmath6 , and @xmath7 respectively .
+ + + keywords : bianchi type i model , dark energy , experimental tests + pacs number : 98.80.es , 98.80-k , 95.36.+x |
discussions on quantum measurement @xcite usually rely on two fundamental elements , the born rule @xcite , telling us how to extract information from the quantum mechanical wave function , and the von neumann @xcite projection postulate , stating that measuring a system observable generates a collapse of the wave function and telling us how to restart the system s unitary evolution after the measurement .
a third most important element is that of the detector s back - action on the system which is at the heart of the projection process when viewed from a microscopic perspective ; much effort has gone into the understanding of the phenomenological von neumann projection in terms of a unitary evolution of the entangled system detector dynamics . here ,
we take a step back and study the charge dynamics , specifically the charge - charge correlator , of a quantum dot ( qd ) as measured by a quantum point contact ( qpc ) in order to understand the impact of the detector s back - action on the time evolution of the correlator .
we study the impact of the back - action in two limiting cases : i ) a weak detector system coupling which we treat perturbatively describing the limit of no back - action and ii ) an intermediate / strong system detector coupling which we describe by a von neumann projection accounting for the limit of strong back - action .
we determine the physically measureable correlators and quantitatively analyze their difference for the case of a quantum dot with a single resonant level .
understanding quantum measurement as the bridge between the quantum- and our classical work is a fascinating and broad topic , ranging from fundamental aspects of the measurement problem @xcite to such practical issues as optimizing the information gain at minimal system invasion @xcite .
much effort has gone into the microscopic understanding of the measurement process and its interrelation with back - action , treating quantum averaged evolutions @xcite , selective system dynamics @xcite , and correlated weak - strong measurements in the form of weak values @xcite .
quantum dots in transport @xcite , isolated double quantum dots @xcite , as well as quantum point contacts @xcite have played a central role in analyzing quantum measurement within the realm of mesoscopic physics . in the present study ,
we focus on the charge dynamics of a quantum dot ( or a localized region of a mesoscopic scatterer in more general terms ) as measured by a nearby qpc detector and determine the time evolution of the average charge and the charge - charge correlator for different strengths of the system detector coupling .
thereby we keep in mind a measurement with single - particle wave packets@xcite incident on the detector at times @xmath0 and @xmath1 with their reflection / transmission through the qpc providing information on the charge state of the quantum dot at the two time instances .
our analysis provides us with two central results : on a technical level , we find that the strong projective measurement can be expressed through a projected charge @xmath2 , the charge at time @xmath3 after a previous projection at time @xmath4 , for which we find a compact expression in terms of the system s scattering matrix . on a physical level , we find that , in spite of the strongly different back - action induced by the first measurement , the two correlators for weak and strong measurements come out qualitatively similar , although quantitative differences remain , of course ; the latter are specifically discussed for the single - resonance level model .
furthermore , we attempt to model the case of a finite constant voltage @xmath5 applied to the qpc detector by considering a sequence of wave packets incident on the qpc , leading to repeated projections which we describe via the projection postulate . increasing the rate of projections , we identify a transition from a regime where the dynamics of the system is dominated by the system s characteristics to a regime where the dynamics is universal and uniquely determined by the measurement . the paper is organized as follows : in sec .
[ sec : setup ] we introduce the model describing our mesoscopic conductor and discuss the different regimes of projective ( intermediate / strong coupling ) and weak measurements , identifying the measurable charge - charge correlator in each of these regimes . in sec.[sec : cc ] , the two charge - charge correlators are calculated for arbitrary scatterers and their difference is discussed quantitatively for the case of a single - level quantum dot . in sec . [
sec : mp ] , the discussion is extended to the case of repeated projective measurements . in order to understand the universal behavior found for fast repeated measurements
, we consider the model of a fluctuating quantum dot level . a summary and conclusions
are given in sec .
[ sec : con ] .
we consider a mesoscopic conductor with a central scattering region ( e.g. , a quantum dot ) , described through its single - particle scattering matrix @xmath6 ( see ref . for a review on the scattering matrix approach to mesoscopic transport ) .
here , we are interested in the dynamics of the charge @xmath7 ( measured in units of electronic charge @xmath8 ) in the scattering region and its modifications when it is subjected to a ( strong ) measurement . in order to measure this dynamics and specifically the associated charge - charge correlator , the scattering region
is capacitively coupled to the quantum point contact ( qpc ) of a detector system , see fig .
[ fig : sysdet ] .
the charge dynamics @xmath7 of the system and the current @xmath9 through the detector then mutually influence one another and we have to analyze the measurement process in order to identify the measurable charge - charge correlator . holding the charge @xmath10 ( here a quantum dot )
capacitively coupled to the central region @xmath11 of the measurement system .
the current @xmath12 in the detector serves as a readout for the charge @xmath10 in the region @xmath13.,width=226 ] we consider a one - dimensional non - interacting system consisting of two half - infinite leads connected through a central region @xmath14 $ ] where particles are scattered by the single - particle potential @xmath15 .
the latter is characterized by the single - particle scattering matrix @xmath16 we make use of the lippmann - schwinger scattering states @xmath17 satisfying @xmath18 with the kinetic part of the hamiltonian @xmath19 , the wavevector @xmath20 and @xmath21 describing a scattering state incoming from the left / right .
we linearize the spectrum around the fermi energy @xmath22 , i.e. , @xmath23 .
the asymptotics @xmath24 of the ( ls ) scattering states is described by the scattering amplitudes , i.e. , @xmath25 spin is trivially accounted for in our non - interacting system and hence we restrict ourselves to spinless particles . as shown in refs . and , the operator @xmath10 describing the charge in the central region @xmath13 ( in units of the electronic charge @xmath8 ) can be expressed through creation ( annihilation ) operators @xmath26 ( @xmath27 ) of the above lippmann - schwinger scattering states via @xmath28 with the notation @xmath29 and @xmath30 .
the matrix elements @xmath31_{a^\prime , a}$ ] are related to the scattering matrix @xmath32 , @xmath33 with @xmath34 and where we have dropped terms of order @xmath35 .
the matrix @xmath36 is equivalent to the density of states matrix expression introduced in ref . and
was used in ref . to express the interaction kernel through the scattering states in a discussion of interacting electron transport
; here , it is used to express the charge through the scattering states , see eq . , and to obtain simple and compact expressions for the projected charge , see eqs . and .
with regard to the latter , we note that the matrix @xmath36 possesses the useful projector property @xmath37 a consequence of the projector property @xmath38 of the charge operator restricted to the single - particle hilbert space ; the latter is easily checked by acting ( twice ) with the charge operator @xmath39 on a single - particle state @xmath40 ( here , @xmath41 denotes the real - space projector on the region @xmath42 ) .
the transport through @xmath13 is implemented by connecting the two semi - infinite leads to two reservoirs at chemical potential @xmath43 and @xmath44 with @xmath45 as described by the steady state density matrix @xmath46 . within this model , we are able to describe local properties of the system , specifically the charge dynamics in the region @xmath13 , through the single - particle scattering matrix @xmath47 which is an asymptotic property of the system .
note that interactions are limited to the system detector coupling , i.e. , we are considering a non - interacting system ; interactions within the system could be taken into account within perturbation theory @xcite . in order to measure the system charge @xmath10 we make use of a capacitively coupled qpc detector @xcite .
the latter is characterized by a step - like transmission characteristic , see fig .
[ fig : meas_reg ] , with a width @xmath48 related to the tunneling time @xmath49 of particles traversing the constriction .
the coupling @xmath50 between the system charge @xmath10 and the charge @xmath51 in the qpc region @xmath11 of the detector will shift the location of the transmission step and the detector current will provide information about the system , while at the same time cause an unavoidable back - action . .
in the case of strong coupling @xmath52 , the change of transmission between @xmath53 ( empty dot ) and @xmath54 ( filled dot ) can be tuned to about unity ( see black dots ) .
( b ) for weak coupling , @xmath55 , the change of transmission ( see black dots ) is small , i.e. , @xmath56 . ] to fix ideas , we restrict the discussion to a single - level quantum dot ( qd ) with two charge states @xmath57 and @xmath58 and focus on the limit where each detector electron probes the dot individually , requiring that the tunneling time @xmath59 is small compared to the typical separation between detector electrons ( alternatively , we can consider a weakly driven detector with bias voltage @xmath60 such that the typical time separation @xmath61 between electrons is large ) . in order for the detector to measure the system it should be fast , @xmath62 , where @xmath63 is a typical system timescale .
in this situation , the system is in a fixed charge state during the tunneling of a detector electron .
if the qd is empty ( state @xmath57 ) , the electron is transmitted with probability @xmath64 , while for a filled qd ( state @xmath58 ) , the transmission probability is @xmath65 .
we then identify two measurement regimes : i ) a strong coupling regime with @xmath66 , see fig .
[ fig : meas_reg](a ) , where we can choose a working point such that @xmath67 and @xmath68 , and ii ) a weak coupling regime @xmath69 , see fig .
[ fig : meas_reg](b ) , where the change in transmission @xmath70 is small . for strong coupling ,
@xmath66 the charge states of the dot and the state of one single qpc electron after scattering [ i.e. , transmitted ( @xmath71 ) or reflected ( @xmath72 ) ] become fully entangled , @xmath73 at time @xmath3 at strong coupling @xmath66 , one single electron @xcite is sent towards the qpc to arrive there at time @xmath3 .
the detection of the transmitted electron determines the charge state of region @xmath42 and projects the system to the corresponding state .
repeating this measurement ( after the system has equilibrated ) allows to determine the average charge @xmath74 .
similarly , in order to measure the charge - charge correlator , the dot charge has to be measured at times @xmath0 and @xmath1 and the measurement has to be repeated after equilibration at fixed time delay @xmath75 in order to find the average charge - charge correlator for times @xmath0 and @xmath1 . in order to describe this correlator
theoretically , we describe the measurement using the projection postulate @xcite : the measurement of the charge operator @xmath10 at @xmath3 projects the system onto a state with an integer number of charges in the central region @xmath13 which we account for by the operator @xmath76 projecting the state of the system at time @xmath3 onto a state with @xmath77 charges in the central region @xmath13 . while the system s state for times @xmath78 is described by the steady - state density matrix @xmath46 , the charge measurement at time @xmath3 projects the density matrix onto well - defined charge states which are expressed through the operator @xmath79 as @xmath80 . then the _
projected _ charge - charge correlator at times @xmath0 and @xmath81 is given by @xmath82 where the collapse of the state at time @xmath0 is accounted for by the projection operators @xmath83 while the measurement at time @xmath1 does not affect the result .
making use of the cyclic property of the trace and @xmath84 = 0 $ ] , we obtain the projected correlator @xmath85 with @xmath86 and the operator @xmath87 describing the charge in the central region at time @xmath1 provided that the state of the system was projected at time @xmath88 , @xmath89 eq
. corresponds to the measurable charge - charge correlator in the regime of strong , projective measurements .
taking into account the spin degree of freedom , eq .
remains unchanged with @xmath90 and @xmath91 ; furthermore , the up- and down spin components of charge are uncorrelated .
at intermediate coupling @xmath92 the scattering of one single electron becomes a probabilistic process ( with finite transmission and reflection probability for both charge states ) such that one electron alone does not provide the information on the dot s state and many electrons are required to probe the dot .
the full information about the dot state is acquired after the passage of @xmath77 electrons when the measured transmitted charge @xmath93 through the qpc can be attributed to a particular dot state , i.e. , the difference @xmath94 has to be larger than the standard deviation @xmath95 of the probabilistic process of charge transmission . at small temperatures ,
the latter is determined by the charge partitioning noise , i.e. , @xmath96 , such that the required number of probe electrons is @xcite @xmath97 for the case where a finite voltage @xmath5 is applied across the qpc detector , we obtain the measurement time @xmath98 with @xmath99 the voltage time . for small temperatures @xmath100
the partitioning noise dominates over the thermal noise .
next , we have to account for the time scale @xmath63 of the dot . for @xmath101 ,
the state of the dot can be determined by probing the dot for some measurement time larger than @xmath102 ; such a measurement is strong and projective . on the other hand ,
the measurement is generically weak if @xmath103 and we have to find an alternative procedure to find the information on the dot s charge state ; we will discuss this weak measurement regime in the next section .
the discussion in this section has been limited to the case of a single - level quantum dot , however , the extension to the case of a more complex system with more charge states is straightforward .
while in such a situation it is not possible to resolve all charge states by one electron alone , a projective measurement involving many electrons or a weak measurement as described below are still possible . at generically weak coupling @xmath104 ( such that @xmath105 ) , the system can not be measured during the system time @xmath106 and we need an alternative measurement scheme . the capacitive coupling between the system and the detector still affects the detector current @xmath107 and thus can be used to learn about the system s charge dynamics @xmath108 .
the detector current is given by the expectation value @xmath109 , where @xmath110 is the current operator in the heisenberg representation and @xmath111 is the steady - state density matrix of the combined system detector setup , @xmath112 .
for a weak system detector coupling , we can calculate the modulation in the detector current perturbatively @xcite ; to lowest order in the coupling strength @xmath113 the result is @xmath114 where @xmath115 is the current operator in the interaction representation ( i.e. , with respect to the coupling hamiltonian ) and @xmath116 . in the expression above
, @xmath117 is the imaginary part of the correlator between the detector current @xmath118 and the detector charge @xmath119 , i.e. , @xmath120 \rangle\rangle/2 $ ] with @xmath121 .
the quantity @xmath122 is a response function of the measurement apparatus @xmath11 and hence a pure detector property , independent of the system .
a good detector is much faster than the dynamics of the system of interest , i.e. , @xmath123 , such that the charge @xmath124 can be treated as slowly varying on the time scale of the detector .
the response function @xmath125 then is effectively given by the zero frequency response function , i.e. , @xmath126 $ ] , such that @xmath127 this relation allows one to determine the time averaged charge expectation value @xmath128 from an experimentally measured time trace of @xmath129 by calculating the time averaged current @xmath130 .
next , we derive an expression for the charge correlator at times @xmath0 and @xmath1 in the weak - coupling regime .
we consider two detectors@xcite @xmath131 and @xmath132 capacitively coupled to the system , i.e. , @xmath133 .
the correlation between the two detectors arises due to the coupling between the two detectors via the system charge .
the measurable time - domain correlator of the detector is the _
irreducible _ symmetrized current - current correlator@xcite @xmath134 ( in contrast to a measurement in the frequency domain , where a non - symmetrized correlator is measured at positive frequencies @xcite ) .
the above symmetrized detector correlator is related to the charge - charge correlator of the dot via @xmath135 as follows from a perturbative analysis in the lowest order of the couplings . in the above expression , @xmath136\rangle\rangle/2\end{aligned}\ ] ]
are the real and imaginary parts of the current - charge correlator in the detectors @xmath137 or @xmath138 representing pure detector response functions . on the other hand , @xmath139 and @xmath140
\rangle\rangle/2 $ ] are the real ( symmetrized ) and imaginary ( anti - symmetrized ) part of the _ irreducible _ charge - charge correlator of the scattering region @xmath42 which we are interested in . again , for good , i.e. , fast @xmath123 , detectors , the system quantities @xmath141 and @xmath142 are slowly varying on the time scales of the detectors and the response functions effectively are given by the zero frequency response functions @xmath143 and @xmath144 .
for the qpc detector with a symmetric scattering potential , the real part of the zero frequency response vanishes @xcite , i.e. , @xmath145 , and we arrive at the simple relation@xcite @xmath146 with @xmath147 for @xmath148 .
hence , for weak coupling and good ( i.e. , fast ) detectors , the measureable quantity is the symmetrized _ irreducible _ correlator @xmath149 which depends only on the time - difference @xmath150 . analyzing the measurement with a single detector @xmath11
, we find that the relation between the current - current and charge - charge correlators is of the form @xcite @xmath151 with @xmath152 describing the intrinsic current fluctuations of the detector current giving rise to a reduced signal - to - noise ratio .
finally , the measurable _ reducible _ charge - charge correlator that can be compared to the projected charge correlator eq .
is given by @xmath153 as for strong coupling , spin can be taken into account by considering the full charge operator @xmath154 in eq . and accounting for the statistical independence of @xmath155 and @xmath156 .
the measurable quantity at weak coupling is the symmetrized charge - charge correlator @xmath157 . making use of eq . expressing the charge @xmath158 in region @xmath13 through the matrix @xmath159 and the density matrix @xmath160
we obtain the result @xmath161 , \nonumber\end{aligned}\ ] ] with the occupation numbers @xmath162 .
below , we will compare this correlator to the projected charge - charge correlator @xmath163 .
starting from the expression for the projected charge - charge correlator , we make use of the poisson formula @xmath164 to rewrite the projected charge in the form @xmath165 using the projector property of @xmath166 , see eq . , we find for the commutators ( we adopt the notation in eq . ) @xmath167 & = ( e^{-i\lambda}-1)\sum_{\gamma ' } a_{\alpha',\gamma'}(t_1 ) \hat{c}_{\gamma'}^\dagger e^{-i\lambda\hat{q}(t_1)},\nonumber\\ [ \hat{c}_{\alpha}^{\phantom{\dagger}},e^{i\lambda\hat{q}(t_1 ) } ] & = e^{i\lambda\hat{q}(t_1)}(e^{i\lambda}-1 ) \sum_{\gamma } \hat{c}_{\gamma}^{\phantom{\dagger}}a_{\gamma,\alpha}(t_1),\end{aligned}\ ] ] and carrying out the integration in over @xmath168 , the projected charge operator can be written as @xmath169 the matrix @xmath170 associated with the projected charge assumes the form ( in lead space ) @xmath171 \label{eq : ap1}\end{aligned}\ ] ] with the matrix @xmath172 given by eq . and @xmath173 .
finally , the projected reducible charge - charge correlator @xmath174 can be expressed in the form @xmath175,\nonumber\end{aligned}\ ] ] where the first contribution corresponds to the product @xmath176 .
the comparison with the symmetrized correlator @xmath157 found in eq .
reveals quite some similarities , with one of the matrices @xmath177 to be replaced by its projected version @xmath178 .
further below , we will analyze this correlator and compare it to the symmetrized correlator @xmath157 . before doing so
, we show that the projected charge matrix elements in eq .
can be simplified considerably . to this end
, we derive a generalized version of eq . which involves different times @xmath3 and @xmath179 , @xmath180 with the matrix elements @xmath181 @xmath182 and @xmath183 the heaviside function .
combining eqs .
and we obtain the projected charge matrix @xmath184 , \label{eq : ap}\end{aligned}\ ] ] where @xmath185 .
note the considerable simplification of this expression as compared to the original formula , where the structure of the region @xmath42 encoded in the matrix elements @xmath186 through the scattering matrices @xmath187 and @xmath188 appeared three times , while now only one matrix @xmath189 remains .
the result eqs . with for the projected charge operator is the main technical result of this paper .
the generalized projector property eq .
is related to the projector property of the charge operator @xmath190 restricted to the single - particle hilbert space . expressing the field operators in @xmath191 ( restricted to the single - particle hilbert space ) through the scattering states , we can express @xmath192 through the projectors onto the incoming and outgoing parts of the scattering states , @xmath193 where @xmath194 with @xmath195 . here , we have used that @xmath196 and have neglected the small overlaps @xmath197 with @xmath198 , we obtain the relation @xmath199,\end{aligned}\ ] ] where we could drop the term @xmath200 since the outcoming component of a scattered particle can not contribute to the charge on the dot at a later time . expressing @xmath192 through the charge matrix @xmath201 ,
see eq . , and using @xmath202 with matrix elements @xmath203 given above , we straightforwardly arrive at eq .. choosing @xmath204 , we multiply eq . by @xmath205 from the right .
the two correlators @xmath157 and @xmath163 are formally different , see eqs . and .
this implies that in the two limiting measurements , a different charge dynamics is detected due to the different back - action of the measurement onto the system . to study the difference between these two correlators beyond a formal level , we focus on the specific example of a single level quantum dot ( modeled by the single resonance level model@xcite ) and investigate the differences between @xmath157 and @xmath163 quantitatively . in the end
, we will draw some conclusions from this analysis for the general situation .
we model a single - level quantum dot scatterer by a single resonance level with a wavevector @xmath206 defining its energy @xmath207 , a width @xmath208 ( in @xmath209-space ) , and a parameter @xmath210 $ ] describing the asymmetry in the coupling to the two leads @xcite .
the unprojected charge matrix elements in eq . assume the simple form @xmath211 \label{eq : asrl}\end{aligned}\ ] ] with @xmath212 and @xmath213 @xmath214 where @xmath215 .
alternatively , the charge operator can be expressed as @xmath216 with @xmath217 , where the operator @xmath218 creates a charge on the resonance level .
this model describes a single spin - degenerate dot level in the non - interacting case with the spin trivially accounted for as noted above ; alternatively , the model applies to the case of a dot with strong coulomb interaction where at most one electron is present on the dot . in the following ,
we restrict ourselves to the discussion of the equilibrium case at zero temperature , i.e. , @xmath219 and @xmath220 ( more precisely , we require that @xmath221 $ ] ) . eq . combined with eq . then provide us with the equilibrium ( steady - state ) charge @xmath222 which is the measured dot occupation both for a strong- and a weak system detector coupling .
[ [ projected - charge . ] ] projected charge .
+ + + + + + + + + + + + + + + + + first , we determine the expectation value of the projected charge operator @xmath223 , see eq . .
the measurement of this quantity involves two charge projections at times @xmath224 and later at @xmath225 , where the outcome of the first measurement is disregarded .
for the single resonance level , the projected charge matrix element eq .
is given by @xmath226\end{aligned}\ ] ] with the projected amplitude @xmath227 @xmath228 .
the expectation value of the projected charge operator then takes the form @xmath229 where we have introduced the dimensionless time @xmath230 and the relative wave - vector @xmath231 . for @xmath232 , @xmath233 , @xmath234 , @xmath235 , and @xmath236 .
( b ) escape dynamics of the projected electron [ hole ] out of the dot described by @xmath237 ( dotted ) [ @xmath238 ( dashed ) ] resulting in the total projected charge dynamics @xmath239 ( full line ) for @xmath240 .
( c ) sketch of two possible outcomes @xmath241 and @xmath242 of the charge projection at @xmath224.,width=226 ] the time dependence of @xmath243 for different positions of the resonance with respect to the fermi level is plotted in fig.[fig : chargeproj ] .
the data is ( particle - hole ) symmetric with respect to the resonance s position relative to the fermi level , with the average projected charges below the fermi level ( @xmath244 ) and symmetrically above ( @xmath245 ) related via @xmath246 this symmetry enforces the projected charge to be equal to the equilibrium value for the resonance aligned with the fermi level , @xmath247 at @xmath248 , see fig.[fig : chargeproj](a ) .
away from the fermi level , the average projected charge at time @xmath225 is suppressed in magnitude with respect to its equilibrium value @xmath249 and returns back as @xmath250 with a typical equilibration time @xmath251 .
consider , for example , a situation where the resonance is located below the fermi level ( @xmath252 ) .
in this situation , it is more likely to observe an electron on the dot and thus one has @xmath253 in this regime .
surprisingly , the second projective measurement at @xmath254 gives an average charge which is _ smaller _ than its equilibrium value ( and correspondingly for @xmath255 ) . in order to understand the relaxation dynamics of @xmath256
it is instructive to analyze the evolution of the entire many - particle state right after the first projection .
consider again the situation where the resonance level is placed below the fermi level .
after the first projection , the dot is either occupied or empty , i.e. , for zero temperature the state of the system is given by @xmath257 where the equilibrium state @xmath258 corresponds to a filled fermi sea , see fig .
[ fig : chargeproj](c ) . making use of the projectors @xmath259 and @xmath260 one can see , that for both measurement outcomes @xmath261 the projection creates a single electron - hole pair in the system .
the hole then screens the excess charge on the dot and gives rise to a friedel - type oscillation of the charge distribution around the dot due to the sharp fermi edge at zero temperature .
assume the first measurement results in an outcome @xmath262 increasing the charge value on the dot above the equilibrium value .
since the projection is local , the accompanying hole is also created near the dot . during the subsequent evolution the charge in the dot can equilibrate in two ways : 1 )
the electron in the dot can tunnel out above the fermi sea , or 2 ) it can tunnel out below the fermi sea and fill the hole state .
the appearance of the second relaxation channel enhances the electron escape rate that gives rise to the reduction of the average charge value below the equilibrium level , see fig .
[ fig : chargeproj](b ) .
the same picture holds for the state @xmath242 with the excess hole in the dot compensated by an excess electron screening the charge outside the dot . finally , the projected charge expectation value @xmath243 is a weighted average of the two processes corresponding to the two alternatives @xmath261 of the first projection , @xmath263 , see fig .
[ fig : chargeproj](b ) .
the creation of the electron hole - pair due to the projection of the charge provides energy to the system .
the average energy of the excited electron - hole pair is given by @xmath264 , @xmath265 [ \delta q^2+\gamma^2]}.\end{aligned}\ ] ] this integral is divergent for the linear dispersion @xmath266 [ the same is true for the quadratic dispersion as well ( with @xmath267 in eq . ) ] , a consequence of the description of the projection as an instantaneous process . in an experimental realization
the measurement involves a finite time and provides a finite amount of energy to the system , hence the particle - hole pair involves a finite energy .
e.g. , for the case of single - electron projection with a strong coulomb coupling @xmath52 the measurement time is given by the tunneling time @xmath268 and the energy exchange between the system and the detector is limited by @xmath113 ; on the other hand , the measurement process should provide the energy for a ` reasonable ' electron - hole pair ( defining one electron or hole in the dot ) given by @xmath269 and a fast measurement with @xmath270 the coulomb energy exceeds the electron - hole energy as @xmath271 . in the case of a multi - electron projection with @xmath55 ,
each individual qpc electron provides a small amount of energy @xmath272 exciting the system to a virtual state and the overall process creates the resulting electron - hole pair in a similar way as the ionization via multi - photon absorption @xcite . [ [ charge - charge - correlator . ] ] charge - charge correlator .
+ + + + + + + + + + + + + + + + + + + + + + + + + when discussing the charge - charge correlator , it is convenient to subtract the asymptotic value @xmath273 and define @xmath274 ( which corresponds to the irreducible correlator ) and @xmath275 . for a single resonance level , the correlator @xmath276 as given by eq . takes the form @xmath277 on the other hand , the projected correlator @xmath278 follows from eq . and reads @xmath279 with the equilibrium charge expectation value @xmath74 given in eq .. due to particle - hole symmetry , both correlators eqs . and remain unchanged when replacing @xmath280 .
furthermore , the difference between the two correlators is given by the first terms and vanishes as @xmath281 ( filled and empty dot , unaltered by the projection ) as well as at the particle - hole symmetric point at half - filling , @xmath282 .
( dashed line ) and @xmath283 ( full line ) as a function of the dimensionless time @xmath284 for different values of the dimensionless resonance position @xmath285 . while the two correlators are similar for small @xmath286 ( see ( a ) @xmath287 and ( b ) @xmath288 ) , for larger @xmath286 , the difference between the two correlators is more pronounced [ ( c ) @xmath289 and ( d ) @xmath290 .
the absolute value of the correlators decreases for increasing @xmath291 such that in the limit @xmath292 both correlators vanish.,width=283 ] analyzing the time dependence of the two correlators , see figs.[fig : qqxf0neu ] ( a)-(d ) , we observe that both correlators become negative at @xmath293 ( cut off at @xmath294 as @xmath295 drops below unity ) indicating the anti - correlation in the system : detecting a particular charge value on the dot at @xmath224 , the observed charge state of the dot after the tunneling time is more likely to be inverse . at large times both correlators show a weak oscillating behavior approaching @xmath296 and the full charge - charge correlator assumes its asymptotic value @xmath297 . comparing the two correlators with one another , we identify two regimes : ( 1 ) for @xmath298 , the system is close to the particle - hole symmetric point and the time - dependence of the two correlators is qualitatively the same with only a small quantitative difference .
( 2 ) when the fermi level is away from the resonance level , the two correlators deviate considerably as the projection enhances the non - equilibrium behavior of the charge at later times . indeed
, the projected correlator shows a larger anti - correlation and more pronounced oscillations than the symmetric one , see fig .
[ fig : sqqmin ] .
( dashed line ) and @xmath299 ( full line ) as a function of the dimensionless resonance position @xmath285 .
the anticorrelation dip is much more pronounced for the projected correlator.,width=226 ] the following general conclusions can be drawn from the above results : if the region @xmath13 is in a well - defined charge state , the two correlators are identical as the projection due to the first charge measurement does not alter the state . for small deviations from a well - defined charge state ,
the difference between the two correlators remains small .
the same is true for a system with particle - hole symmetry , as discussed for the half - filled dot above .
if the system is in a superposition of different charge states with similar weights ( but away from a particle - hole symmetric point ) , the projection has a pronounced impact on the measurement outcome and the two correlators @xmath300 and @xmath301 can deviate considerably with specific details depending on the scattering matrix of the system .
so far , we have discussed two regimes of measuring the time - resolved charge - charge correlator : i ) by projections ( strong measurements ) at times @xmath0 and @xmath1 , or ii ) by a weak measurement where the charge - charge correlator is obtained from a deconvolution of the detector current - current correlator and the detector response function . while
in the latter case , the detector current may be applied continuously , for the strong measurement it is crucial to measure the charge at two times in a strong manner and allow for an unperturbed evolution in - between .
let us compare these two regimes to a typical experimental situation @xcite : in these experiments , a fixed voltage @xmath5 is applied at the detector which thus is constantly monitoring the charge of the system ; at the same time , the clear steps visible in the detector - current traces indicate a strong measurement . given the constant monitoring of the dot charge , the system is always perturbed . furthermore , while our analysis above assumes a fully coherent evolution of the system , the coupling to the environment leads to a further decoherence beyond the one introduced by the measurement .
our model of a projective measurement of the charge - charge correlator @xmath302 with an initial projection at time @xmath0 and a final projection at time @xmath1 is inappropriate to describe these measurements . in the following ,
we formulate the continuous strong measurement discussed above and derive the appropriate charge - charge correlator . following the general spirit of the paper ,
we model the continuous strong measurement by repeated projective measurements of the system .
we introduce the formalism of multiple projective measurements and present the results for the multiple projected charge expectation value as well as the multiple projected charge - charge correlator .
we identify a critical measurement rate where the dynamics of the charge changes from a system property to a regime where it is dominated by the measurement , becoming universal in the limit of a high projection rate . in the end
, we discuss the rate at which we should project the system in order to model a realistic continuous measurement and compare to experiments .
starting from the situation where we project the density matrix @xmath46 once , see sec .
[ sec : cc - strong ] above , we account for @xmath303 projections at times @xmath304 via the multiply projected density matrix @xmath305 \hat{\rho}_0 \tilde{\mathcal{t}}\bigl[\prod_{i=1}^{n}\hat{p}_{n_i}(t_i)\bigr]\end{aligned}\ ] ] with the time - ordering operator @xmath306 and the anti - time - ordering operator @xmath307 .
the charge expectation at time @xmath308 after @xmath303 previous projections at times @xmath304 is given by the expression @xmath309
\hat{\rho}_0 \tilde{\mathcal{t}}\bigl[\prod_{i=1}^{n}\hat{p}_{n_i}(t_i)\bigr ] \bigr\}\nonumber\\ & = \sum_{n_1,\ldots , n_n } \text{tr}\bigl\ { \tilde{\mathcal{t } } \bigl [ \prod_{i=1}^{n}\hat{p}_{n_i}(t_i ) \bigr ] \hat{q}(t)\mathcal{t } \bigl [ \prod_{i=1}^{n}\hat{p}_{n_i}(t_i ) \bigr ] \hat{\rho}_0 \bigr\ } \nonumber\nonumber\\ & \equiv\langle\hat{q}^{mp}(t|\{t_j\})\rangle,\end{aligned}\ ] ] where we have used again the trace property @xmath310 and we have introduced the multiply projected charge operator @xmath311 with the projection times @xmath312 .
the latter can be expressed as @xmath313 with matrix elements @xmath314 . to obtain these matrix elements ,
the projection operators can be taken into account iteratively , i.e. , @xmath315 . making use of eq . and the property @xmath316 with @xmath317
, the resulting matrix element can be simplified to @xmath318_{k^\prime p^\prime } { \bf a}_{p^\prime , p}(t ) \bigl[\pi(t_{i+1})-\pi(t_{i})\bigr]_{pk}\nonumber\\ & \quad+ \bar{\pi}_{k^\prime p^\prime}(t_n ) { \bf a}_{p^\prime , p}(t ) \bar{\pi}_{pk}(t_n ) \bigr ] .
\label{eq : amultpr}\end{aligned}\ ] ] the relevant timescale in the description of ( multiple ) projections of the mean charge is the equilibration time @xmath319 ( for @xmath320 ) . for multiple projections ,
there are two regimes : if the time between subsequent projections @xmath321 exceeds the equilibration time , @xmath322 , the projections are independent , see fig.[fig : chargemultproj ] ( dashed line ) for a plot of the corresponding multiply projected charge @xmath323 . on the other hand , for short intervalls @xmath324 ,
the system is not in equilibrium when the next projection occurs ; multiple charge projections with small equal time steps @xmath325 then drive the system to a new steady state , see fig.[fig : chargemultproj ] ( thin line ) .
returns to the equilibrium value after a characteristic time scale @xmath326 ( upper thick line , we choose @xmath327 ) .
for multiple projections ( indicated by arrows ) with time steps @xmath325 , we find two different behaviors for the charge @xmath328 : for @xmath329 , the dot returns to the equilibrium before the next projection ( dashed line , @xmath330 ) , while this is not the case when @xmath331 and the system is driven to a new steady state ( upper thin line , @xmath332 ) . for faster projections ( lower thin line , @xmath333 ) , the occupation value is driven towards 1/2 ( with the limit @xmath334 indicated by the lower thick line ) .
, width=264 ] when the fermi energy resides above the resonance level ( @xmath335 ) , the new steady state corresponds to a lower occupation of the dot state than in equilibrium , while for a resonance level above the fermi energy ( @xmath336 ) , the dot occupation is increased .
frequent projection pushes the steady state occupation towards 1/2 , see fig.[fig : chargemultproj ] . for @xmath337 , the steady state occupation ( at the times of projection )
is given by @xmath338 i.e. , the dot occupation is 1/2 up to a small correction .
we can attribute this universal behavior ( independent of applied chemical potentials ) to an effective ` broadening ' of the level over the width @xmath339 in energy as this width becomes much larger than the distance @xmath340 of the resonance level to the fermi level , the dot is filled and emptied with equal probability .
let us compare these results and in particular the appearance of a universal regime with a charge expectation value 1/2 to the experiments by gustavsson _
_ @xcite ; as it turns out , this experiment is not dominated by the measurement but rather by temperature and no universal behavior is expected .
the experimental setup consists of a qpc detector which is measuring the charge states @xmath341 or @xmath342 of a capacitively coupled quantum dot .
the qpc detector is driven with a constant voltage @xmath343 producing a current @xmath344 which corresponds to electrons impinging on the qpc at a rate @xmath345 . from the current steps ,
we conclude that the transmissions corresponding to the two charge states are given by @xmath346 and @xmath347 , implying a coupling @xmath348 . making use of eq .
, we obtain that @xmath349 electrons determine the dot state . for each data point , the qpc current is integrated during @xmath350 such that @xmath351 electrons pass through the qpc , which is far larger than @xmath352 . with @xmath353
every data point then provides the information of the actual charge state of the dot .
the system lifetime derives from the step widths ( of order @xmath354 ms ) of the measurement tracks and amounts to @xmath355 .
we model the effect of the continuous measurement by repeated projections on the time scale @xmath356 , that provides an effective level broadening @xmath357 .
this value is far below the temperature scale @xcite @xmath358 ( at @xmath359 ) .
alternatively , the system is drive - dominated when @xmath360 , see ref .. next , we investigate the effect of multiple projections on the charge - charge correlator .
we consider the correlator between charges at times @xmath0 and @xmath3 with multiple projections at times @xmath361 in - between , assuming a free evolution of the system at times prior to the first measurement at @xmath0 .
the corresponding correlator is given by @xmath362 \hat{q}(t_1)\hat{\rho}_0 \tilde{\mathcal{t}}\bigl[\prod_{i=1}^{n}\hat{p}_{n_i}(t_i)\bigr ] \bigr\ } \nonumber\\ & = \text{tr}\bigl\ { \hat{q}^{mp}(t|\{t_j\ } ) \hat{q}(t_1)\hat{\rho}_0 \bigr\},\end{aligned}\ ] ] where we have used the trace property and the definition of the multiply projected charge operator @xmath311 .
making use of the matrix elements @xmath363 in eq .
, the charge - charge correlator can be expressed as @xmath364 , \nonumber\end{aligned}\ ] ] where the first contribution is @xmath365 . comparing this result to the case of a single projection discussed before ,
, we recognize that the multiple projections enter through the matrix elements @xmath363 replacing the ( singly projected ) matrix element @xmath366 .
for the single resonance level in equilibrium ( @xmath367 ) the charge matrix element assumes the simple form in eq . . in fig .
[ fig : ppqmultproj ] we compare the projected correlator @xmath368 with the multiply projected correlator @xmath369 with equal time separations @xmath370 between projections ; we consider the cases @xmath371 and @xmath372 .
all correlators start out at @xmath373 with the value @xmath74 of the average charge eq .. the correlator @xmath374 then undergoes a pronounced anti - correlation dip and approaches @xmath375 for @xmath250 ( as the two charge states at @xmath376 and @xmath3 become uncorrelated ) .
repeated projections in between drive the correlator @xmath377 towards @xmath378 as the measurement induces an additional time dependence . in the limit of frequent projections ,
the projected charge expectation value @xmath379 approaches 1/2 and hence the asymptotic value of the correlator is given by @xmath380 .
overall , the result shows that only small times @xmath381 probe the system dynamics , while larger times become dominated by the measurement . for multiple projections at @xmath382 : single projected correlator ( thick line ) ,
multiply projected correlator ( dashed line , with @xmath383 and dots indicating the projections ) , and the limiting behavior for @xmath384 ( thin line).,width=264 ] we have seen that repeated projections with a time separation between subsequent projections approaching the heisenberg time @xmath385 start dominating the dynamic evolution of the system and the average occupation number of the dot approaches the universal value @xmath386 , irrespective of the position of the dot level with respect to the fermi level , see eq . .
here , we show how this formal result of the von neumann projection postulate can be understood in terms of a measurement - induced fluctuation of the dot level . as before
, we assume that the state of the dot is measured by a capacitively coupled qpc , see fig .
[ fig : osc_level ] .
we assume single electrons passing through the qpc with a frequency @xmath387 , e.g. , due to an applied voltage @xmath5 corresponding to @xmath388 , which corresponds to an incoming qpc current @xmath389 .
an electron passing through the qpc interacts typically during a time @xmath390 with the dot .
the interaction strength is given by the coulomb energy @xmath113 . in our model , we take into account the effect of this current on the dot level @xmath391 via a classical electrostatic potential of the form @xmath392 with an effective coupling strength @xmath393 accounting for the typical interaction time @xmath390 . in order to describe the system , we define the time - dependent hamiltonian coupled to two leads with states filled up to the fermi energy @xmath22 is capacitively interacting ( with interaction strength @xmath113 ) with the electrons passing through the qpc with frequency @xmath387 .
we take this interaction into account via an effective time - dependent level energy @xmath394.,width=264 ] @xmath395 , \nonumber\end{aligned}\ ] ] where @xmath396 , @xmath397 are the annihilation operators for the left and right reservoirs , @xmath398 is the annihilation operator corresponding to the localized dot state , @xmath399 are the partial widths of the level , @xmath400 is the linearized dispersion of the lead electrons , and @xmath394 is the fluctuating energy of the dot . solving the heisenberg equation of motion , we obtain @xmath401 where @xmath402 is the resonance width and @xmath403 . taking @xmath404 and @xmath373 , the steady - state dot occupation is given by @xmath405 which in equilibrium , i.e. , @xmath406 , simplifies to @xmath407 the phase @xmath408 is of the form @xmath409 and using the expansion @xmath410 , we can rewrite the expression above as @xmath411 with @xmath412 the bessel functions of the first kind . assuming frequent projections , @xmath413 , and a narrow resonance , @xmath414 , we can calculate @xmath415 to leading order , where only terms @xmath416 contribute , and obtain @xmath417,\ ] ] where the identity @xmath418 ^ 2 + 2\sum^{\infty}_{\nu=1}[j_{\nu}(x)]^2 $ ] has been used . making use of @xmath419
, we notice that for @xmath52 , the steady - state charge @xmath415 approaches the asymptotic value @xmath420 . hence , assuming a strong system - detector coupling with each electron realizing a strong projective measurement , we reproduce the case of frequent projective measurements discussed before , see eq . . in order to reproduce the case of a projection by @xmath77 electrons within a time @xmath102
, we consider a periodic oscillation @xmath421 of the dot level with frequency @xmath422 and effective coulomb strength @xmath423 .
while this periodicity is artificially introduced here , it could be realized by switching the qpc current on and off repeatedly with this period .
adapting the calculation above , we obtain in the limit of fast projection @xmath424 an average charge @xmath425 $ ] , such that for @xmath426 the system is driven to the universal regime as in the case of single electron projection discussed above . within this treatment ,
it is the randomization of the dot level which leads to equal probabilities to find the level either empty or filled and which is responsible for the universal filling factor 1/2 .
we have studied the charge and charge - charge correlator of a quantum dot @xmath13 ( a localized region @xmath13 of an arbitrary mesoscopic scatterer in more general terms ) . making use of the single - particle scattering matrix of the scatterer @xmath427
, we expressed the charge operator @xmath158 for @xmath13 through the single - particle scattering matrix . in order to measure this charge ,
the system has been coupled to a detector ; the measurement then acts back on the dot s occupation and therefore the charge dynamics depends on the coupling to the detector .
we have focussed on a qpc detector which is capacitively coupled to the charge in region @xmath13 .
we have analyzed different measurement regimes and identified two particular cases : at strong dot detector coupling @xmath66 , one single electron already can perform a strong measurement and project the system ; such a strong / projective measurement is still possible even at intermediate coupling @xmath428 where many electrons accumulated over a measurement time @xmath429 are needed to project the dot - state , provided that @xmath102 remains below the system time @xmath63 where the charge on the dot changes . in our phenomenological approach ,
we have described such strong measurements with the help of the von neumann projection postulate ( maximal back action ) and the measured charge - charge correlator is the projected quantity @xmath430 involving the projected charge operator @xmath223 .
when the coupling is weak , @xmath104 and @xmath431 the measurement is in the weak regime .
we have analyzed this case in lowest - order perturbation theory ( with vanishing back action ) and have identified the symmetrized irreducible correlator @xmath432 as the measurable quantity .
the difference between the strong and weak measurement regimes consists in the different degree the system becomes entangled with the detector during the systems s dynamical time . in a strong measurement ,
the projection of the detector degree of freedom implies a projection of the system s state , allowing to exclude the detector from the consideration . for the weak measurement ,
the system becomes only weakly entangled with the detector and almost preserves its coherent evolution ; the detector s projection ( involving a quasi - classical variable @xcite ) then acts only as a weak perturbation on the system s dynamics .
although the results for strong and weak measurements look quite different on a formal level , we find a qualitative correspondence when analyzing the expressions for a single - level quantum dot .
the projected charge correlator involves the projected charge operator @xmath223 providing the charge at time @xmath3 after a projection at an earlier time @xmath224 , see eq . .
a simple expression for this quantity ( that could also be carried over to the case of multiply repeated projections ) was obtained by making use of the analytic structure of the scattering matrix @xmath47 .
the evolution of this projected charge , in particular , its initial decay in modulus and subsequent return to the equilibrium value , could be understood in terms of the generation of a particle - hole excitation in the nearby leads through the projection . in the limit of frequent projections ,
we have identified a transition from a regime where the dot dynamics is determined by the characteristics of the system to a regime where it is uniquely determined by the detector with universal outcome . in order to reproduce these results within a unitary treatment ,
we have modeled the projective measurements by a strongly fluctuating dot level .
the back - action of a measurement naturally manifests itself in the time - correlator of the measured quantity , as the second measurement probes both the dynamics of the system as well as the back - action on the system originating from the first measurement . furthermore , the strength of the back - action increases from zero in a weak measurement to a maximum in a strong measurement described by the von neumann projection scheme .
our analysis of time - correlators then provides insights on the range of back - action effects in a simple model setup .
we thank simon gustavsson and oded zilberberg for illuminating discussions and acknowledge financial support from the swiss national science foundation through the national center of competence in research on quantum science and technology ( qsit ) , the pauli center for theoretical studies at eth zurich , and the rfbr grant no . 14 - 02 - 01287 .
s. gustavsson , r. leturcq , b. simovic , r.schleser , t. ihn , p. studerus , k. ensslin , d.c .
driscoll , and a.c.gossard , phys .
lett . * 96 * , 076605 ( 2006 ) ; s. gustavsson , r.leturcq , b. simovic , r. schleser , p. studerus , t. ihn , k. ensslin , d.c .
driscoll , and a.c .
gossard , phys .
rev . b * 74 * , 195305 ( 2006 ) .
averin , in _ exploring the quantum - classical frontier : recent advances in macroscopic and mesoscopic quantum phenomena _ , edited by j.r .
friedman and s. han ( nova science publishers , new york , 2003 ) , pp .
447 - 470 . | we study the charge dynamics of a quantum dot as measured by a nearby quantum point contact probing the dot via individual single - particle wave packets .
we contrast the two limiting cases of weak and strong system detector coupling exerting vanishing and strong back - action on the system and analyze the resulting differences in the charge - charge correlator . extending the study to multiple projective measurements modelling a continuous strong measurement , we identify a transition from a charge dynamics dominated by the system s properties to a universal dynamics governed by the measurement . |
multiple - description coding ( mdc ) aims at creating separate descriptions individually capable of reproducing a source to a specified accuracy and when combined being able to refine each other .
traditionally quantizer based mdc schemes consider only two descriptions @xcite . among the few vector quantizer based approaches which consider more than two descriptions are @xcite . in @xcite closed form expressions for the design of lattice vector quantizers
are given for the symmetric case where all packet - loss probabilities and side entropies are equal .
-channel system .
descriptions are encoded at an entropy of @xmath1 , @xmath2 .
the erasure channel either transmits the @xmath3th description errorless or not at all.,width=340 ] in @xcite iterative vector quantizer design algorithms are proposed for the asymmetric case where packet - loss probabilities and side entropies are allowed to be unequal .
in this paper we consider the asymmetric case for an arbitrary number of descriptions , where the @xmath3th description is encoded at an entropy of @xmath1 , for @xmath2 , see fig .
[ fig : nchannel ] .
the total rate is then given by the sum of the entropies of the individual descriptions .
due to the asymmetry , the total distortion depends not only on how many descriptions are received ( as is the case in the symmetric situation @xcite ) , but also on _ which _ descriptions make it to the decoder .
we derive analytical expressions for the central and side quantizers which , under high - resolution assumptions , minimize the _ expected distortion _ at the receiver subject to entropy constraints on the total rate .
in contrast to @xcite our design allows for simple adaptation of our quantizers to changing source - channel characteristics and entropy constraints , effectively avoiding iterative quantizer design procedures .
let @xmath4 be an arbitrary i.i.d .
source and let @xmath5 be a real lattice with voronoi regions @xmath6 , given by @xmath7 where @xmath8 is a realization of @xmath9 and we define @xmath10 , where @xmath11 denotes vector transposition .
we consider one central lattice ( central quantizer ) @xmath12 and several sublattices ( side quantizers ) @xmath13 , where @xmath2 and @xmath14 , is the number of descriptions .
the trivial case @xmath15 leads to a single - description system , where we would simply use one central quantizer and no side quantizers .
we assume that sublattices are geometrically similar to @xmath12 , i.e. they can be obtained from @xmath12 by applying change of scales , rotations and possible reflections .
the sublattice index @xmath16 , n_i\in \mathbb{z}^+$ ] , of the @xmath3th sublattice @xmath13 describes the volume @xmath17 of a sublattice cell relative to the volume @xmath18 of a central lattice cell .
the volume @xmath17 of the @xmath3th sublattice cell is then given by @xmath19 . in the design of the index assignment map
, we make use of a product lattice @xmath20 , which is simply a sublattice of index @xmath21 $ ] . to simplify the design of the index assignment map we assume sublattices are clean @xcite
, specifically we require that no points of @xmath12 lies on the boundaries of the voronoi regions of @xmath22 .
a source vector @xmath8 is quantized to the nearest reconstruction point @xmath23 in the central lattice @xmath12 .
hereafter follows index assignments ( mappings ) , which uniquely map all @xmath23 s to reconstruction points in each of the sublattices @xmath13 .
this mapping is done through a labeling function @xmath24 , and we denote the individual component functions of @xmath24 by @xmath25 . in other words , the injective map @xmath24 that maps @xmath12 into @xmath26 , is given by @xmath27 where @xmath28 and @xmath2 .
each @xmath29-tuple @xmath30 is used only once when labeling points in @xmath12 in order to make sure that @xmath23 can be recovered unambiguously when all @xmath29 descriptions are received .
since lattices are infinite arrays of points , we adopt the procedure used in @xcite and construct a shift invariant labeling function , so only a finite number of points must be labeled .
we generalize the approach of @xcite and construct a product lattice @xmath22 which has @xmath31 central lattice points and @xmath32 sublattice points from the @xmath3th sublattice in each of its voronoi regions .
the voronoi regions @xmath33 of the product lattice @xmath22 are all similar so by labeling only central lattice points within one voronoi region of @xmath22 , the rest of the central lattice points may be labeled simply by translating this voronoi region throughout @xmath34 . without loss of generality ,
we let @xmath35 and by construction we let @xmath22 be a geometrical similar and clean sublattice of @xmath13 as well as @xmath12 . with this choice of @xmath22
, we only label central lattice points within @xmath36 , which is the voronoi region of @xmath22 around origo . with this
we get the following shift invariant property @xmath37 for all @xmath38 and all @xmath39 . using standard high - resolution assumptions for lattice quantizers @xcite
, the expected central distortion can be expressed as @xmath40 where @xmath41 is the normalized second moment of inertia @xcite of the central quantizer and it can be shown that the side distortion for the @xmath3th description is given by @xcite @xmath42 the minimum entropy @xmath43 needed to achieve the central distortion @xmath44 is given by @xcite @xmath45 where @xmath46 is the component - wise differential entropy of the source .
the side entropies are given by @xcite @xmath47
the index assignment is done by a labeling function @xmath24 , that maps central lattice points to sublattice points .
an optimal assignment minimizes the expected distortion when @xmath48 descriptions are received and is invertible so the central quantizer can be used when all descriptions are received . at the receiving side ,
@xmath4 is reconstructed to a quality that is determined by the received descriptions .
if no descriptions are received we reconstruct using the expected value , @xmath49 $ ] , and if all @xmath29 descriptions are received we reconstruct using the inverse map outlined above , hence obtaining the quality of the central quantizer . in all other cases ,
we reconstruct to the average , under high - resolution assumptions , is constant within the region where elements of a @xmath29-tuple are located .
] of the received descriptions .
there are in general several ways of receiving @xmath50 out of @xmath29 descriptions .
let @xmath51 denote an index set consisting of all possible @xmath50 combinations out of @xmath52 so that @xmath53 .
we denote an element of @xmath51 by @xmath54 .
the complement @xmath55 of @xmath56 denotes the @xmath57 indices not in @xmath56 , i.e. @xmath58 .
we will use the notation @xmath59 to indicate the set of all @xmath60 that contains the index @xmath3 , i.e. , @xmath61 and similarly @xmath62 .
furthermore , let @xmath63 be the packet - loss probability for the @xmath3th description and , consequently , let @xmath64 be the probability that the @xmath3th description is received .
finally , let @xmath65 , @xmath66 , @xmath67 and @xmath68 .
for example for @xmath69 and @xmath70 we have @xmath71 and hence @xmath72 . upon reception of any @xmath50 out of @xmath29 descriptions
we reconstruct @xmath9 as @xmath73 where the resulting distortion can be written similar to ( [ eq : di ] ) , e.g. if descriptions @xmath3 and @xmath74 are received , the norm in ( [ eq : di ] ) should read @xmath75 .
it follows that the expected distortion is given by @xmath76 where @xmath77 and the two special cases @xmath78 are given by @xmath79\prod_{i=0}^{k-1}p_i$ ] and @xmath80 . from ( [ eq : expdist ] ) we see that the distortion @xmath81 may be split into two terms , one describing the distortion occurring when the central quantizer is used on the source , and one that describes the distortion due to the index assignment .
an optimal index assignment minimizes the second term in ( [ eq : expdist ] ) for all possible combinations of descriptions .
we can rewrite this term using the following theorem [ theo : sums ] for any @xmath82 we have @xmath83 see @xcite .
the cost functional to be minimized can then be written as @xmath84 we minimize this cost functional subject to a constraint on the sum of the side entropies .
we remark here that the side entropies depend solely on @xmath18 and @xmath85 and as such not on the particular choice of @xmath29-tuples . in other words , for fixed @xmath85 s and a fixed @xmath18 , the index assignment problem is solved if ( [ eq : costfunctional2 ] ) is minimized .
the problem of choosing @xmath18 and @xmath85 such that certain entropy constraints are not violated is independent of the assignment problem and deferred to section [ sec : optq ] .
the first term in ( [ eq : costfunctional2 ] ) describes the distance from a central lattice point to the weighted centroid of its associated @xmath29-tuple .
the second term describes the weighted sum of pairwise squared distances ( wspsd ) between elements of the @xmath29-tuples .
it can be shown , c.f .
proposition [ prop : growthriemann2 ] , that , under a high - resolution assumption , the second term in ( [ eq : costfunctional2 ] ) is dominant , from which we conclude that in order to minimize ( [ eq : costfunctional2 ] ) we must use @xmath29-tuples with the smallest wspsd .
these @xmath29-tuples are then assigned to central lattice points in such a way , that the first term in ( [ eq : costfunctional2 ] ) is minimized .
this problem can be posed and solved as a linear assignment problem @xcite . to obtain @xmath29-tuples
we center a region @xmath86 around all sublattice points @xmath87 , and construct @xmath29-tuples by combining sublattice points from the other sublattices ( i.e. @xmath88 ) within @xmath89 in all possible ways and select the ones that minimize ( [ eq : costfunctional2 ] ) . for each @xmath90
it is possible to construct @xmath91 different @xmath29-tuples , where @xmath92 is the number of sublattice points from the @xmath3th sublattice within the region @xmath86 .
this gives a total of @xmath93 @xmath29-tuples when all @xmath94 are used .
let @xmath95 be the volume of @xmath86 .
since @xmath96 and we need @xmath97 @xmath29-tuples for each @xmath98 , we see that @xmath99 so in order to obtain at least @xmath97 @xmath29-tuples , the volume of @xmath86 must satisfy @xmath100 for the symmetric case , i.e. @xmath101 , @xmath102 , we have @xmath103 , which is in agreement with the results obtained in @xcite . by centering @xmath86 around each @xmath104 , we make sure that the map @xmath24 is shift - invariant .
however , this also means that all @xmath29-tuples have their first coordinate ( i.e. @xmath105 ) inside @xmath36 . to be optimal
this restriction must be removed which is easily done by considering all cosets of each @xmath29-tuple .
the coset of a fixed @xmath29-tuple , say @xmath106 where @xmath90 , is given by @xmath107 , for all @xmath38 .
the @xmath29-tuples in a coset are distinct modulo @xmath22 and by making sure that only one member from each coset is used , the shift - invariance property is preserved .
before we outline the design procedure for constructing an optimal index assignment we remark that in order to minimize the wspsd between a fixed @xmath108 and the set of points @xmath109 it is required that @xmath86 forms a sphere centered at @xmath108 . 1 . center a sphere @xmath86 at each @xmath90 and construct all possible @xmath29-tuples @xmath110 where @xmath111 and @xmath112 .
notice that all @xmath29-tuples have their first coordinate ( @xmath105 ) inside @xmath36 and they are therefore shift - invariant .
make @xmath86 large enough so at least @xmath97 distinct @xmath29-tuples are found for each @xmath105 .
2 . construct cosets of each @xmath29-tuple .
the @xmath31 central lattice points in @xmath113 must now be matched to distinct @xmath29-tuples .
this is a standard linear assignment problem @xcite where only one member from each coset is ( allowed to be ) matched to a central lattice point in @xmath36 . as observed in @xcite ,
having equality in ( [ eq : vtilde ] ) , i.e. using the minimum @xmath95 , will not minimize the wspsd .
instead a slightly larger region must be used . for
the _ practical _ construction of the @xmath29-tuples this is not a problem , since we simply use e.g. twice as large a region as needed and let the linear assignment algorithm choose the optimal @xmath29-tuples .
however , in order to _ theoretically _ describe the performance of the quantizers we need to know the optimal @xmath95 . in @xcite an expansion factor @xmath114
was introduced and used to describe how much @xmath86 had to be expanded from the theoretical lower bound ( [ eq : vtilde ] ) , to make sure that the @xmath97 optimal @xmath29-tuples could be constructed by combining sublattice points within the region @xmath86 . adopting this approach leads to @xmath115 where e.g. for the two - dimensional case @xmath116@xcite .
analytical expressions for @xmath114 are given in @xcite .
in this section we derive high - resolution approximations for the expected distortion . however , we first introduce proposition [ prop : riemann2 ] which relates the sum of distances between pairs of sublattice points to @xmath117 , the dimensionless normalized second - moment of an @xmath0-dimensional sphere .
hereafter follows proposition [ prop : growthriemann2 ] which determines the dominating term in the expression for the expected distortion .
[ prop : riemann2 ] for @xmath118 and @xmath119 , we have for any pair of sublattices , @xmath120 , @xmath121 & \approx \psi^{2/l}\nu^{2/l } g(s_l ) n_\pi\prod_{m=0}^{k-1}n_m^{2/l(k-1)}. \end{split}\ ] ] let @xmath122 , i.e.the set of @xmath31 sublattice points @xmath123 associated with the @xmath31 central lattice points within @xmath36 .
furthermore let @xmath124 be the set of unique elements of @xmath125 , where @xmath126 .
finally , let @xmath127 so that @xmath128 contains all the elements @xmath129 which are in the @xmath29-tuples that also contains a specific @xmath123 .
let @xmath130 be the set of unique elements
. for sublattice @xmath13 and @xmath131 we have @xmath132 observe that each @xmath133 is used @xmath134 times , so given @xmath135 , we have @xmath136 since @xmath137 .
hence , with @xmath138 , we have @xmath139 & \approx n_i\nu_j \psi^{2/l } \nu^{2/l}g(s_l)\prod_{m=0}^{k-1}n_m^{2/l(k-1 ) } , \end{split}\ ] ] which is independent of @xmath108 , so that @xmath140 which completes the proof . [ prop : growthriemann2 ] let @xmath141 be chosen such that @xmath142 for all @xmath143 . for @xmath118 and @xmath144
we have @xmath145 see @xcite .
the expected distortion ( [ eq : expdist ] ) can by use of theorem [ theo : sums ] be written as @xmath146 by use of propositions [ prop : riemann2 ] , [ prop : growthriemann2 ] and eq .
( [ eq : d0 g ] ) it follows that ( [ eq : da ] ) can be written as @xmath147 where @xmath148 depends on @xmath29 and @xmath50 and is given by @xmath149 the total expected distortion is obtained by summing over @xmath50 including the cases where @xmath150 and @xmath151 , @xmath152\prod_{i=0}^{k-1}p_i \\ & \quad + \psi^{2/l}\nu^{2/l } g(s_l)\prod_{m=0}^{k-1}n_m^{2/l(k-1)}\hat{\beta } , \end{split}\ ] ] where @xmath153 and @xmath154 . using ( [ eq : rc ] ) and ( [ eq : ri ] ) we can write the expected distortion as a function of entropies , which leads to @xmath155\prod_{i=0}^{k-1}p_i\\ & + \psi^{2/l } \hat{\beta } g(s_l)2^{2(h(x)-r_c)}2^{\frac{2k}{k-1}\left(r_c - \frac{1}{k}\sum_{i=0}^{k-1}r_i\right)}. \end{split}\ ] ]
in this section we consider the situation where the total bit budget is constrained , i.e. we find the optimal scaling factors , @xmath85 and @xmath18 , subject to entropy constraints on the sum of the side entropies @xmath156 , where @xmath157 is the target entropy .
we also find the optimal bit - distribution among the @xmath29 descriptions .
first we observe from ( [ eq : daopt ] ) that the expected distortion depends upon the _ sum _ of the side entropies and not the individual side entropies . in order to be optimal
it is necessary to achieve equality in the entropy constraint , i.e. @xmath158 . from ( [ eq : ri ] ) we have @xmath159 which can be rewritten as @xmath160 where @xmath161 is constant for fixed target and differential entropies .
writing ( [ eq : tmptau ] ) as @xmath162 and inserting in ( [ eq : adistopt ] ) leads to @xmath163\prod_{i=0}^{k-1}p_i \\ & \quad + \psi^{2/l}\nu^{-2/l(k-1)}\tau_*^{2/l(k-1 ) } g(s_l)\hat{\beta}. \end{split}\ ] ] the optimal @xmath18 is found by differentiating ( [ eq : edisttmp ] ) w.r.t .
@xmath18 , equating to zero and solving for @xmath18 , which leads to @xmath164 at this point we still need to find expressions for the optimal @xmath1 ( or , equivalently , optimal @xmath85 given @xmath18 ) .
let @xmath165 , where @xmath166 , hence @xmath158 .
from ( [ eq : ri ] ) we have @xmath167 which can be rewritten as @xmath168 where , after inserting the optimal @xmath18 from ( [ eq : optnurt ] ) we obtain an expression for the optimal index value @xmath85 , that is @xmath169 it follows from ( [ eq : ri ] ) that @xmath170 so that @xmath171 . in addition , since rates must be positive , we obtain the following inequalities @xmath172 hence , the individual side entropies @xmath173 can be arbitrarily chosen as long as they satisfy ( [ eq : ai ] ) and @xmath174 .
to verify theoretical results we present in this section experimental results obtained by using @xmath175 two - dimensional zero - mean unit - variance gaussian source vectors .
[ fig : a2k4_perf ] shows the theoretical expected distortion ( [ eq : adistopt ] ) and the numerical expected distortion obtained for @xmath176 descriptions when using the @xmath177 quantizer @xcite at a total entropy @xmath178 bits / dimension . in this setup
we have @xmath179 and packet - loss probabilities are fixed at @xmath180 except for @xmath181 which is varied in the range @xmath182\%$ ] .
as @xmath181 is varied we update @xmath18 according to ( [ eq : optnurt ] ) and arbitrarily pick the index values @xmath85 such that @xmath183 .
however , index values are restricted to a certain set of integers @xcite and the side entropies might therefore not sum exactly to @xmath157 . to make sure
the target entropy is met with equality we then rescale @xmath18 as @xmath184 .
we see from fig . [
fig : a2k4_perf ] a good correspondence between the theoretically and numerically obtained results .
[ fig : a2k4perf ]
this research is supported by the technology foundation stw , applied science division of nwo and the technology programme of the ministry of economics affairs .
v. a. vaishampayan , n. j. a. sloane , and s. d. servetto , `` multiple - description vector quantization with lattice codebooks : design and analysis , '' _ ieee trans .
_ , vol .
47 , no . 5 , pp .
1718 1734 , july 2001 .
j. stergaard , j. jensen , and r. heusdens , `` entropy constrained multiple description lattice vector quantization , '' in _ proc .
acoust . , speech , and signal proc .
_ , vol . 4 , may 2004 , pp
601 604 . | we present analytical expressions for optimal entropy - constrained multiple - description lattice vector quantizers which , under high - resolutions assumptions , minimize the expected distortion for given packet - loss probabilities .
we consider the asymmetric case where packet - loss probabilities and side entropies are allowed to be unequal and find optimal quantizers for any number of descriptions in any dimension . we show that the normalized second moments of the side - quantizers are given by that of an @xmath0-dimensional sphere independent of the choice of lattices .
furthermore , we show that the optimal bit - distribution among the descriptions is not unique .
in fact , within certain limits , bits can be arbitrarily distributed . |
neutrino oscillations have provided the first experimental manifestation of flavour violation in the lepton sector , fuelling the need to consider extensions of the standard model ( sm ) that succeed in explaining the smallness of neutrino masses and the observed pattern of their mixings @xcite .
in addition to the many facilities dedicated to study neutral leptons , there is currently a great experimental effort to search for signals of flavour violation in the charged lepton sector ( clfv ) , since such an observation would provide clear evidence for the existence of new physics beyond the sm ( trivially extended to accommodate massive neutrinos ) .
the quest for the origin of the underlying mechanism of flavour violation in the lepton sector has been actively pursued in recent years , becoming even more challenging as the meg experiment is continually improving the sensitivity to @xmath3 decays @xcite , thus opening the door for a possible measurement ( observation ) in the very near future .
the current bounds on other radiative decays ( i.e. @xmath12 ) , or three - body decays ( @xmath13 ) are already impressive @xcite , and are expected to be further improved in the future .
supersymmetric ( susy ) extensions of the sm offer new sources of cp and flavour violation , in both quark and lepton sectors . given the strong experimental constraints , especially on cp and flavour violating observables involving the strongly interacting sector , phenomenological analyses in general favour the so - called `` flavour - blind '' mechanisms of susy breaking , where universality of the soft breaking terms is assumed at some high energy scale : in these constrained scenarios , the only sources of flavour violation ( fv ) are the quark and charged lepton yukawa couplings . in order to accommodate current neutrino data , mechanisms of neutrino mass generation , such as the seesaw ( in its different realisations - for a review of the latter , see for instance @xcite ) , are often implemented in the framework of ( constrained ) susy models : in the case of the so - called `` susy - seesaw '' , radiatively induced flavour violation in the slepton sector @xcite can provide sizable contributions to clfv observables .
the latter have been extensively studied , both at high- and low - energies , over the past years ( see e.g. @xcite ) .
flavour violation can be also incorporated in a more phenomenological approach , where at low - energies new sources of fv are present in the soft susy breaking terms . however , these are severely constrained by a large number of observables ( see , e.g. @xcite and references therein ) .
in addition to the above mentioned rare lepton decays , leptonic and semi - leptonic meson decays also offer a rich testing ground for clfv . here
we will be particularly interested in leptonic @xmath14 decays , which ( as is also the case of leptonic @xmath15 decays ) constitute very good probes of violation of lepton universality .
the potential of these observables , especially regarding susy extensions of the sm , was firstly noticed in @xcite , and later investigated in greater detail in @xcite . by themselves , these decays are heavily hampered by hadronic uncertainties and , in order to reduce the latter ( and render these decays an efficient probe of new physics ) , one usually considers the ratio @xmath16\right)}{\gamma\left(k^{+}\rightarrow\mu^{+}\nu \left[\gamma\right]\right)}\,,\label{eq : rk : def}\ ] ] since in this case the hadronic uncertainties cancel to a very good approximation . as a consequence , the sm prediction can be computed with high precision @xcite .
the most recent analysis has provided the following value @xcite : @xmath17 on the experimental side , the na62 collaboration has recently obtained very stringent bounds @xcite : @xmath18 which should be compared with the sm prediction ( eq . ( [ eq : cirigliano:2007xi ] ) ) . in order to do so , it is often useful to introduce the following parametrisation , @xmath19 where @xmath20 is a quantity denoting potential contributions arising from scenarios of new physics ( np ) . comparing the theoretical sm prediction to the current bounds ( i.e. , eqs .
( [ eq : cirigliano:2007xi ] , [ eq : rk : na62 ] ) ) , one verifies that observation is compatible with the sm ( at 1@xmath21 ) for @xmath22 previous analyses have investigated supersymmetric contributions to @xmath2 in different frameworks , as for instance low - energy susy extensions of the sm ( i.e. the unconstrained minimal supersymmetric standard model ( mssm ) ) @xcite , or non - minimal grand unified models ( where higher dimensional terms contribute to fermion masses ) @xcite .
these studies have also considered the interplay of @xmath2 with other important low - energy flavour observables , magnetic and electric lepton moments and potential implications for leptonic cp violation .
distinct computations , based on an approximate parametrisation of flavour violating effects - the mass insertion approximation ( mia ) @xcite - allowed to establish that susy lfv contributions can induce large contributions to the breaking of lepton universality , as parametrised by @xmath20 .
the dominant fv contributions are in general associated to charged - higgs mediated processes , being enhanced due to non - holomorphic effects - the so - called `` hrs '' mechanism @xcite - , and require flavour violation in the @xmath23 block of the charged slepton mass matrix .
it is important to notice that these higgs contributions have been known to have an impact on numerous observables , and can become especially relevant for the large @xmath24 regime @xcite . in the present work
, we re - evaluate the potential of a broad class of supersymmetric extensions of the sm to saturate the current measurement of @xmath2 .
contrary to previous studies , we conduct a full computation of the one - loop corrections to the @xmath1 vertex , taking into account the important contributions from non - holormophic effective higgs - mediated interactions .
when possible we establish a bridge between our results and approximate analytical expressions in the literature , and we stress the potential enhancements to the total susy contributions .
in our numerical analysis we re - investigate the prospects regarding @xmath2 of a constrained mssm onto which several seesaw realisations are embedded ( type i @xcite and ii @xcite , as well as the inverse seesaw @xcite ) , also briefly addressing @xmath25@xmath26 symmetric models @xcite .
we then consider more relaxed scenarios , such as non - universal higgs mass ( nuhm ) models at high - scale ( which are known to enhance this class of observables @xcite due to potentially lighter charged higgs boson masses ) , and discuss the general prospects of unconstrained low - energy susy models . in all cases ,
we revisit the @xmath2 observable in the light of new experimental data : in addition to lhc bounds gev . ] on the sparticle spectrum @xcite and a number of low - energy flavour - related bounds @xcite , we implement the very recent lhcb results concerning the br(@xmath27 ) @xcite .
as we discuss here , the increasing tension with low - energy observables , in particular with @xmath6 , precludes sizable susy contributions to @xmath2 even in the context of otherwise favoured candidate models as is the case of semi - constrained and unconstrained susy models .
this document is organised as follows . section [ sec : rk : formulae ] is devoted to the computation of the 1-loop mssm prediction for @xmath2 .
we compare our ( full ) result to the approximations in the literature by means of the mass insertion approximation ( among other simplifications ) , and discuss the dominant sources of flavour violation , and the implications to other observables .
our results for a number of models are collected in section [ sec : res ] . further discussion and concluding remarks are given in section [ sec : concs ] . in the appendices ,
we detail the computation of the renormalised charged lepton - neutrino - charged higgs vertex , and summarise the key features of two supersymmetric seesaw realisations ( types i and ii ) used in the numerical analysis .
in the sm , the decay widths of pseudoscalar mesons into light leptons are given by @xmath28 where @xmath29 denotes @xmath30 or @xmath4 mesons , with mass @xmath31 and decay constant @xmath32 , and where @xmath33 is the fermi constant , @xmath34 the lepton mass and @xmath35 the corresponding cabibbo - kobayashi - maskawa ( ckm ) matrix element .
these decays are helicity suppressed ( as can be seen from the factor @xmath36 in eq .
( [ eq : sm : pdecays ] ) ) , and the prediction for their amplitude is thus hampered by the hadronic uncertainties in the meson decay constants .
as mentioned in the introduction , ratios of these amplitudes are independent of @xmath32 to a very good approximation , and the sm prediction can then be computed very precisely .
concerning the kaon decay ratio @xmath2 , the sm prediction ( inclusive of internal bremsstrahlung radiation ) is @xcite @xmath37 where @xmath38 is a small electromagnetic correction accounting for internal bremsstrahlung and structure - dependent effects ( @xmath39 @xcite ) . [ cols="^ " , ] as can be seen from both panels of fig .
[ fig : deltarcomparisoncuts ] , values @xmath40 could be obtainable , in agreement with refs @xcite .
however , the situation is substantially altered when one takes into account the current experimental bounds on @xmath4 decays ( @xmath6 and @xmath27 ) and @xmath41 . as is manifest from the left panel of fig .
[ fig : deltarcomparisoncuts ] , once experimental bounds - other than @xmath6 - are imposed , one could in principle have @xmath42 ; however , taking into account the limits from br(@xmath6 ) , one is now led to @xmath43 .
a few comments are in order regarding the impact of the different low - energy bounds from radiative tau decays and @xmath4-physics observables .
firstly , let us consider the @xmath41 decay : although directly depending on @xmath44 , its amplitude is ( roughly ) suppressed by the fourth power of the average susy scale , @xmath45 . as can be seen from eqs .
( [ eq : deltar : approx : x ] , [ eq : deltar : approx ] ) , @xmath20 only depends on the charged higgs mass - if the latter is assumed to be an ew scale parameter , @xmath20 will be thus independent of @xmath45 in these unconstrained models . as such
, it is possible to evade the @xmath41 bound by increasing the soft susy masses , and this can indeed be seen from the right - hand panel of fig .
[ fig : deltarcomparisoncuts ] , where a number of `` blue '' points are found to lie below the br(@xmath41 ) bound .
secondly , the @xmath27 decay can be a severe constraint regarding the susy contributions to @xmath20 in the case of constrained models ( see , e.g. , figs .
[ fig : cmssm : seesaw ] and [ fig : cmssm : survey ] ) .
we notice that @xmath27 is approximately proportional to @xmath46 ( see for instance @xcite ) while @xmath20 shows no such dependence : thus a regime of small trilinear couplings easily allows evade the @xmath27 bounds .
finally , let us discuss the @xmath6 bounds .
notice that this is a process essentially identical to the charged kaon decays at the origin of the @xmath2 ratio ( the only difference being that the @xmath47 meson is to be replaced by a @xmath48 and the @xmath49/@xmath50 in the decay products by a kinematically allowed @xmath51 ) , and hence its tree - level decay width can be inferred from eqs .
( [ eq : sm : pdecays ] ) and ( [ eq : kaon : gamma : smsusy ] ) . due to a negative interference between the sm and the mssm contributions , given by the term proportional to @xmath52 in eq .
( [ eq : kaon : gamma : smsusy ] ) , regimes of low @xmath53 lead to excessively small values of @xmath6 ( below the experimental bound ) , effectively setting a lower bound for for @xmath54 ( see right panel of fig . [
fig : nuhm : deltar ] , in relation to the discussion of nuhm models ) . in turn
, this excludes regimes of @xmath53 associated to sizable values of @xmath20 , as is clear from the comparison of the `` blue '' and `` green '' regions of the left panel of fig .
[ fig : deltarcomparisoncuts ] . in summary , we conclude that saturating the experimental bound on @xmath2 clearly proves to be extremely difficult ( if not impossible ) , even in the unconstrained mssm , especially in view of the stringent constraints from @xmath6 .
in this work we have revisited supersymmetric contributions to @xmath55@xmath56 , considering the potential of a broad class of constrained susy models to saturate the current measurement of @xmath2 .
we based our analysis in a full computation of the one - loop corrections to the @xmath1 vertex ; we have also derived ( when possible ) illustrative analytical approximations , which in addition to offering a more transparent understanding of the rle of the different parameters , also allow to establish a bridge between our results and previous ones in the literature .
our analysis further revisited the @xmath2 observable in the light of new experimental data , arising from flavour physics as well as from collider searches .
we numerically evaluated the contributions to @xmath2 arising in the context of different minimal supergravity inspired models which account for observed neutrino data , further considering the possibility of accommodating a near future observation of a @xmath3 decay .
as expected from the ( mostly ) @xmath57 nature of the radiatively induced charged lepton flavour violation , type i and ii seesaw mechanisms implemented in the cmssm provide minimal contributions to @xmath2 , thus implying that such cmssm susy seesaws can not saturate the present value for @xmath20 .
we then considered unified models where the flavour - conserving hypothesis on the @xmath23 slepton sector is relaxed by allowing a non - vanishing @xmath44 ( @xmath58 sector ) . in all models ,
special attention was given to experimental constraints , especially four observables which turn out to play a particularly relevant rle : the recent interval for the lightest neutral higgs boson mass provided by the cms and atlas collaborations , br(@xmath27 ) , br(@xmath6 ) and br(@xmath59 ) .
these last two exhibit a dependence on @xmath53 ( @xmath6 ) and on @xmath44 ( @xmath41 ) similar to that of @xmath20 .
the susy contributions to @xmath20 are thus maximised in a regime in which @xmath53 and @xmath44 are such that the experimental limits for @xmath6 and @xmath41 are simultaneously saturated ; in this regime one must then accommodate the bounds on other observables , such as @xmath60 and br(@xmath27 ) . for a minimal deviation from a pure cmssm scenario allowing for non - vanishing values of @xmath44
, we can have values for @xmath20 at most of the order of @xmath61 .
in fact , the requirement of having a higgs boson mass of 125 - 126 gev is much more constraining on the cmssm parameter space than , for instance @xmath27 ( which is sub - dominant , and can be overcome by variations of the trilinear coupling , @xmath62 ) . in order to have @xmath63
, one must significantly increase @xmath44 so to marginally overlap the regions of @xmath64 gev , while still in agreement with @xmath41 .
models where the charged higgs mass can be significantly lowered , as is the case of nuhm models , allow to increase the susy contributions to @xmath20 , which can be as large as @xmath65 ( larger values being precluded due to @xmath6 decay constraints ) .
more general models , as the unconstrained mssm realised at low - energies , offer more degrees of freedom , and the possibility to better accommodate / evade the different experimental constraints . in the unconstrained mssm ,
one can find values of @xmath20 one order of magnitude larger , @xmath66 . again
, any further augmentation is precluded due to incompatibility with the bounds on @xmath6 .
however @xmath66 still remains one order of magnitude shy of the current experimental sensitivity to @xmath2 , and also substantially lower than some of the values previously found in the literature .
as such , if susy is indeed discovered , and unless there is significant progress in the experimental sensitivity to @xmath2 , it seems unlikely that the contributions to @xmath2 of the susy models studied here will be testable in the near future . on the other hand , any near - future measurement of @xmath20 larger than @xmath67 would unambiguously point towards a scenario different than those here addressed ( msugra - like seesaw , nuhm and the phenomenological mssm ) .
it should be kept in mind that the analysis presented here focused on the impact of lfv interactions .
should the discrepancy between the sm and experimental observations turn out to be much smaller than @xmath65 , a more detailed approach and evaluation will then be necessary .
r.m.f . is thankful for the hospitality of the lpc clermont - ferrand .
the work of r.m.f has been supported by _
fundao para a cincia e a tecnologia _ through the fellowship sfrh / bd/47795/2008 .
and j. c. r. also acknowledge the financial support from the eu network grant unilhc pitn - ga-2009 - 237920 and from _ fundao para a cincia e a tecnologia _ grants cftp - fct unit 777 , cern / fp/83503/2008 and ptdc / fis/102120/2008 .
a. m. t. acknowledges partial support from the european union fp7 itn - invisibles ( marie curie actions , pitn- ga-2011 - 289442 ) .
in what follows we detail the computation leading to eqs .
( [ eq : epsilondelta]-[eq : delta : def ] ) , and we further refer to @xcite for a similar analysis . as expected , loop effects contribute to both kinetic and mass terms of charged leptons as well as to the @xmath68 vertex : @xmath69 & \!\ !
+ i\ , \overline{\nu}_{l}\left(\mathbf{1}+\eta_{l}^{\nu}\right ) \slashed{\partial}\nu_{l}- \left[\overline{\ell}_{l } \left(m^{l0}+\eta_{m}^{\ell}\right)\ell_{r}+\textrm{h.c.}\right]\nonumber \\[+1 mm ] & \!\ ! + \left[\overline{\nu}_{l}\left(2^{3/4}g_{f}^{1/2}\,\tan\beta\ , m^{l0}+\eta^{h}\right)\ell_{r}h^{+}+\textrm{h.c.}\right]\,.\end{aligned}\ ] ]
here @xmath70 denotes the bare charged lepton mass and the @xmath71 s correspond to loop contributions to the various terms .
the ( new ) kinetic terms can be recast into a canonical form by means of unitary rotations of the fields ( @xmath72 , @xmath73 , @xmath74 ) , which are then renormalised by diagonal transformations ( @xmath75 , @xmath76 , @xmath77 ) : @xmath78 two unitary rotation matrices ( @xmath79 , @xmath80 ) are further required to diagonalise the charged lepton mass matrix , and one finally has @xmath81 in the new basis , the mass terms now read @xmath82\ , r_{r}^{\ell}\,\ell_{r}+\textrm{h.c.}\ , .
\end{aligned}\ ] ] the above equation relates the unknown parameter @xmath70 with the physical mass matrix @xmath83 . using the latter to rewrite the @xmath68 vertex one finds @xmath84 where @xmath85 to one - loop order
, this exact expression simplifies to @xmath86+\eta^{h}\,.\end{aligned}\ ] ] the expressions for the @xmath71 s can be computed from the relevant feynman diagrams ( assuming zero external momenta ) : @xmath87 -\left(4\pi\right)^{2}\hskip -1 mm & \left(\eta_{r}^{\ell}\right)_{ij}= \ , n_{i\alpha\beta}^{l\left(\ell\right)}n_{j\alpha\beta}^{l\left(\ell\right ) * } b_{1}\left(0,m_{\chi_{\alpha}^{0}}^{2},m_{\widetilde{\ell}_{\beta}}^{2}\right ) \nonumber\\ + & c_{i\alpha\beta}^{l\left(\ell\right)}c_{j\alpha\beta}^{l\left(\ell\right ) * } b_{1}\left(0,m_{\chi_{\alpha}^{\pm}}^{2},m_{\widetilde{\nu}_{\beta}}^{2}\right ) , \label{eq : app : etaellr}\\[+1 mm ] -\left(4\pi\right)^{2}\hskip -1 mm & \left(\eta_{l}^{\ell}\right)_{ij}= \ , n_{i\alpha\beta}^{r\left(\ell\right)}n_{j\alpha\beta}^{r\left(\ell\right ) * } b_{1}\left(0,m_{\chi_{\alpha}^{0}}^{2},m_{\widetilde{\ell}_{\beta}}^{2}\right ) \nonumber\\ + & c_{i\alpha\beta}^{r\left(\ell\right)}c_{j\alpha\beta}^{r\left(\ell\right ) * } b_{1}\left(0,m_{\chi_{\alpha}^{\pm}}^{2},m_{\widetilde{\nu}_{\beta}}^{2 } \right)\!,\label{eq : app : etaelll}\\[+1 mm ] -\left(4\pi\right)^{2}\hskip -1 mm & \left(\eta_{l}^{\nu}\right)_{ij}= \ , n_{i\alpha\beta}^{r\left(\nu\right)}n_{j\alpha\beta}^{r\left(\nu\right ) * } b_{1}\left(0,m_{\chi_{\alpha}^{0}}^{2},m_{\widetilde{\nu}_{\beta}}^{2}\right ) \nonumber\\ + & c_{i\alpha\beta}^{r\left(\nu\right)}c_{j\alpha\beta}^{r\left(\nu\right ) * } b_{1}\left(0,m_{\chi_{\alpha}^{\pm}}^{2},m_{\widetilde{\ell}_{\beta}}^{2}\right ) , \label{eq : app : etanul}\\[+1 mm ] -\left(4\pi\right)^{2}\hskip -1 mm & \left(\eta^{h}\right)_{ij}= \ , c_{i\beta\gamma}^{r\left(\nu\right)}n_{j\alpha\gamma}^{l\left(\ell\right ) * } \left[d_{\beta\alpha2}^{l\left(s^{+}\right)*}m_{\chi_{\alpha}^{0 } } m_{\chi_{\beta}^{\pm}}\right.\nonumber\\ & \left.\hskip 25 mm c_{0}\left(0,0,0,m_{\chi_{\alpha}^{0}}^{2},m_{\chi_{\beta}^{\pm}}^{2},m_{\widetilde{\ell}_{\gamma}}^{2}\right)\right.\nonumber \\ + & \left .
d_{\beta\alpha2}^{r\left(s^{+}\right ) * } dc_{00}\left(0,0,0,m_{\chi_{\alpha}^{0}}^{2},m_{\chi_{\beta}^{\pm}}^{2},m_{\widetilde{\ell}_{\gamma}}^{2 } \right)\right]\nonumber \\ & + \ , n_{i\alpha\gamma}^{r\left(\nu\right ) } c_{j\beta\gamma}^{l\left(\ell\right)*}\left[d_{\beta\alpha2}^{l\left(s^{+}\right ) * } m_{\chi_{\alpha}^{0}}m_{\chi_{\beta}^{\pm } } \right.\nonumber\\ & \left.\hskip 30 mm c_{0}\left(0,0,0,m_{\chi_{\alpha}^{0}}^{2},m_{\chi_{\beta}^{\pm}}^{2},m_{\widetilde{\nu}_{\gamma}}^{2}\right)\right.\nonumber \\ & \left.+\ , d_{\beta\alpha2}^{r\left(s^{+}\right ) * } dc_{00}\left(0,0,0,m_{\chi_{\alpha}^{0}}^{2},m_{\chi_{\beta}^{\pm}}^{2},m_{\widetilde{\nu}_{\gamma}}^{2}\right)\right]\nonumber \\ & + \ , n_{i\alpha\beta}^{r\left(\nu\right ) } n_{j\alpha\gamma}^{l\left(\ell\right ) * } g_{2\gamma\beta}^{\left(s^{+}\widetilde{\ell}\widetilde{\nu}^{*}\right ) } m_{\chi_{\gamma}^{0}}\nonumber\\ & \hskip 20 mm c_{0}\left(0,0,0,m_{\widetilde{\ell}_{\gamma}}^{2},m_{\widetilde{\nu}_{\beta}}^{2},m_{\chi_{\alpha}^{0}}^{2}\right)\,,\label{eq : app : etah}\end{aligned}\ ] ] with @xmath88 denoting the usual loop integral functions @xmath89\,,\label{eq : app : b1}\\ & c_{0}\left(0,0,0,x , y , z\right)=\,\frac{xy\log\frac{x}{y}+yz\log\frac{y}{z}+zx\log\frac{z}{x}}{\left(x - y\right)\left(y - z\right)\left(z - x\right)}\,,\\ & dc_{00}\left(0,0,0,x , y , z\right)=\,\delta_{\varepsilon}+1\nonumber\\ & \hskip -2 mm + \frac{x^{2}\left(y - z\right)\log\frac{x}{\mu^{2}}+y^{2}\left(z - x\right)\log\frac{y}{\mu^{2}}+z^{2}\left(x - y\right)\log\frac{z}{\mu^{2}}}{\left(x - y\right)\left(y - z\right)\left(z - x\right)}\,.\end{aligned}\ ] ] here @xmath90 , @xmath50 is the regularisation parameter and @xmath91 . for the couplings notation we followed @xcite . the comparison of the above expressions with the corresponding ones derived in ref .
@xcite , reveals a fair agreement ; we nevertheless notice that the neutralino and chargino masses are absent from the analogous of eq .
( [ eq : app : etaellm ] ) , and that the order of the arguments of @xmath92 in eqs .
( [ eq : app : etaellr ] , [ eq : app : etaelll ] , [ eq : app : etanul ] ) appears reversed .
moreover , we find small discrepancies ( which can not be accounted by the distinct notations used ) in the expressions for @xmath93 and @xmath94 , cf .
( [ eq : app : etaellm ] ) and eq .
( [ eq : app : etah ] ) , respectively .
in its different realisations , the seesaw mechanism offers one of the most appealing explanations for the smallness of neutrino masses and the pattern of neutrino mixing angles . moreover ,
when embedded in the framework of susy models - the so - called susy seesaw - the seesaw offers the interesting feature that flavour violation in the neutrino sector ( encoded in non - diagonal neutrino yukawa couplings ) can radiatively induce flavour violation in the slepton sector at low - energies @xcite , leading to potentially sizable contributions to a large array of observables . in what follows we
briefly summarise the most relevant features of different realisations of the seesaw mechanism .
in particular , we will consider `` high - scale '' seesaws , i.e. , where the additional states are assumed to be much heavier than the electroweak scale ( in association with large values of the corresponding couplings ) . in a type
i susy seesaw , the mssm superfield content is extended by three right - handed majorana neutrino superfields .
the lepton superpotential is thus extended as @xmath95 where , and without loss of generality , one can work in a basis where both @xmath96 and @xmath97 are diagonal ( @xmath98 , @xmath99 ) .
the relevant slepton soft - breaking terms are now @xmath100 should this be embedded into a cmssm , then the additional soft breaking parameters would also obey universality conditions at the gut scale , @xmath101 and @xmath102 . in this case , the light neutrino masses are given by @xmath103 with @xmath104 ( @xmath105 being the vacuum expectation values ( vevs ) of the neutral higgs scalars , @xmath106 , with @xmath107 gev ) , and where @xmath108 corresponds to the masses of the heavy right - handed neutrino eigenstates .
the light neutrino matrix @xmath109 is diagonalized by the @xmath110 as @xmath111 .
a convenient means of parametrising the neutrino yukawa couplings , while at the same time allowing to accommodate the experimental data , is given by the casas - ibarra parametrisation @xcite , which reads at the seesaw scale , @xmath97 , @xmath112 in the above , @xmath26 is a complex orthogonal @xmath113 matrix that encodes the possible mixings involving the right - handed neutrinos , in addition to those of the low - energy sector ( i.e. @xmath110 ) and which can be parametrised in terms of three complex angles @xmath114 @xmath115 . in our analysis
, we assumed degenerate right - handed neutrino masses and real parameters , so that the results are effectively independent of the choice of the @xmath114 . even under universality conditions at the gut scale ,
the non - trivial flavour structure of @xmath116 will induce ( through the running from @xmath117 down to the seesaw scale , @xmath97 ) flavour mixing in the otherwise approximately flavour conserving soft - susy breaking terms . in particular , there will be radiatively induced flavour mixing in the slepton mass matrices , manifest in the @xmath57 and @xmath118 blocks of the @xmath119 slepton mass matrix ; an analytical estimation using the leading order ( llog ) approximation leads to the following corrections to the slepton mass terms : @xmath120 the amount of flavour violation is encoded in the matrix elements @xmath121 of eq .
( [ eq : lfv : llog ] ) . the implementation of a type ii susy seesaw model requires the addition of at least two su(2 ) triplet superfields @xcite .
should one aim at preserving gauge coupling unification , then complete su(5 ) multiplets must be added to the mssm content . under the sm gauge group ,
the * 15 * decomposes as @xmath122 , where @xmath123 , @xmath124 and @xmath125 . in the su(5 ) broken phase ( below the gut scale ) , the superpotential contains the following terms : where we have omitted flavour indices for simplicity ( for shortness we will not detail the soft breaking lagrangian here , see e.g. @xcite ) . after having integrated out the heavy fields ,
the effective neutrino mass matrix then reads @xmath127 as occurs in the type i seesaw , lfv entries in the charged slepton mass matrix are radiatively induced , and are proportional to the combination @xmath128 @xcite ; for example , the @xmath57 block reads @xmath129 f. p. an _ et al . _
[ daya - bay collaboration ] , arxiv:1203.1669 [ hep - ex ] ; j. k. ahn _
[ reno collaboration ] , arxiv:1204.0626 [ hep - ex ] ; p. adamson _ et al . _
[ minos collaboration ] , arxiv:1202.2772 [ hep - ex ] .
p. minkowski , phys .
b * 67 * ( 1977 ) 421 ; m. gell - mann , p. ramond and r. slansky , in _ complex spinors and unified theories _ eds .
p. van .
nieuwenhuizen and d. z. freedman , _ supergravity _ ( north - holland , amsterdam , 1979 ) , p.315 [ print-80 - 0576 ( cern ) ] ; t. yanagida , in _ proceedings of the workshop on the unified theory and the baryon number in the universe _ ,
eds . o. sawada and a. sugamoto ( kek , tsukuba , 1979 ) , p.95 ; s. l. glashow , in _ quarks and leptons _
, eds . m. lvy _ et al . _ ( plenum press , new york , 1980 ) , p.687 ; r. n. mohapatra and g. senjanovi , phys . rev .
* 44 * ( 1980 ) 912 .
r. barbieri , d. v. nanopolous , g. morchio and f. strocchi , phys .
b * 90 * ( 1980 ) 91 ; r. e. marshak and r. n. mohapatra , _ invited talk given at orbis scientiae , coral gables , fla . ,
14 - 17 , 1980 _ , vpi - hep-80/02 ; t. p. cheng and l. f. li , phys .
d * 22 * ( 1980 ) 2860 ; m. magg and c. wetterich , phys .
b * 94 * ( 1980 ) 61 ; g. lazarides , q. shafi and c. wetterich , nucl .
b181 ( 1981 ) 287 ; j. schechter and j. w. f. valle , phys .
d * 22 * ( 1980 ) 2227 ; r. n. mohapatra and g. senjanovi , phys .
d * 23 * ( 1981 ) 165 .
j. c. pati and a. salam , phys .
d * 10 * ( 1974 ) 275 [ erratum - ibid .
d * 11 * ( 1975 ) 703 ] ; r. n. mohapatra and j. c. pati , phys .
d * 11 * ( 1975 ) 2558 ; g. senjanovic and r. n. mohapatra , phys .
d * 12 * ( 1975 ) 1502 ; c. s. aulakh , k. benakli and g. senjanovic , phys .
* 79 * ( 1997 ) 2188 [ arxiv : hep - ph/9703434 ] .
g. aad _ et al .
_ [ atlas collaboration ] , phys .
b * 710 * , 49 ( 2012 ) [ arxiv:1202.1408 [ hep - ex ] ] ; s. chatrchyan _ et al .
_ [ cms collaboration ] , phys .
b * 710 * ( 2012 ) 26 [ arxiv:1202.1488 [ hep - ex ] ] ; s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 716 * ( 2012 ) 30 [ arxiv:1207.7235 [ hep - ex ] ] ; g. aad _ et al . _
[ atlas collaboration ] , phys .
b * 716 * ( 2012 ) 1 [ arxiv:1207.7214 [ hep - ex ] ] .
g. aad _ et al . _ [ atlas collaboration ] , arxiv:1203.5763 [ hep - ex ] ; g. aad _ et al . _ [ atlas collaboration ] , arxiv:1202.4847 [ hep - ex ] ; s. a. koay and c. collaboration , arxiv:1202.1000 [ hep - ex ] ; g. aad _ et al . _ [ atlas collaboration ] , arxiv:1111.4116 [ hep - ex ] ; g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 710 * , 67 ( 2012 ) [ arxiv:1109.6572 [ hep - ex ] ] ; g. aad _ et al .
_ [ atlas collaboration ] , arxiv:1107.0561 [ hep - ex ] ; g. aad _ et al .
_ [ atlas collaboration ] , eur .
j. c * 71 * ( 2011 ) 1682 [ arxiv:1103.6214 [ hep - ex ] ] ; g. aad _ et al .
_ [ atlas collaboration ] , eur .
j. c * 71 * ( 2011 ) 1647 [ arxiv:1103.6208 [ hep - ex ] ] ; g. aad _ et al .
_ [ atlas collaboration ] , phys .
b * 701 * ( 2011 ) 398 [ arxiv:1103.4344 [ hep - ex ] ] ; g. aad _ et al . _
[ atlas collaboration ] , phys .
b * 701 * ( 2011 ) 1 [ arxiv:1103.1984 [ hep - ex ] ] ; j. b. g. da costa _ et al .
_ [ atlas collaboration ] , phys .
b * 701 * ( 2011 ) 186 [ arxiv:1102.5290 [ hep - ex ] ] ; g. aad _ et al .
_ [ atlas collaboration ] , phys .
* 106 * ( 2011 ) 131802 [ arxiv:1102.2357 [ hep - ex ] ] ; s. chatrchyan _ et al . _
[ cms collaboration ] , cern - ph - ep-2011 - 138 ; s. chatrchyan _ et al . _
[ cms collaboration ] , jhep * 1108 * ( 2011 ) 156 [ arxiv:1107.1870 [ hep - ex ] ] ; s. chatrchyan _ et al . _
[ cms collaboration ] , arxiv:1107.1279 [ hep - ex ] ; s. chatrchyan _ et al .
_ [ cms collaboration ] , jhep * 1108 * ( 2011 ) 155 [ arxiv:1106.4503 [ hep - ex ] ] ; s. chatrchyan _ et al . _ [ cms collaboration ] , jhep * 1107 * ( 2011 ) 113 [ arxiv:1106.3272 [ hep - ex ] ] ; s. chatrchyan _ et al .
_ [ cms collaboration ] , arxiv:1106.0933 [ hep - ex ] ; s. chatrchyan _ et al .
_ [ cms collaboration ] , jhep * 1106 * ( 2011 ) 093 [ arxiv:1105.3152 [ hep - ex ] ] ; s. chatrchyan _ et al . _
[ cms collaboration ] , jhep * 1106 * ( 2011 ) 077 [ arxiv:1104.3168 [ hep - ex ] ] ; s. chatrchyan _ et al .
_ [ cms collaboration ] , jhep * 1106 * ( 2011 ) 026 [ arxiv:1103.1348 [ hep - ex ] ] ; s. chatrchyan _ et al . _ [ cms collaboration ] , phys .
* 106 * ( 2011 ) 211802 [ arxiv:1103.0953 [ hep - ex ] ] ; v. khachatryan _ et al . _
[ cms collaboration ] , phys .
b * 698 * ( 2011 ) 196 [ arxiv:1101.1628 [ hep - ex ] ] ; v. khachatryan _ et al .
_ [ cms collaboration ] , phys .
* 106 * ( 2011 ) 011801 [ arxiv:1011.5861 [ hep - ex ] ]
. j. k. parry , nucl .
b * 760 * ( 2007 ) 38 [ hep - ph/0510305 ] ; j. r. ellis , j. s. lee and a. pilaftsis , phys .
d * 76 * ( 2007 ) 115011 [ arxiv:0708.2079 [ hep - ph ] ] .
g. isidori and p. paradisi , phys .
b * 639 * ( 2006 ) 499 [ hep - ph/0605012 ] .
w. porod , comput .
* 153 * ( 2003 ) 275 [ arxiv : hep - ph/0301101 ] .
m. hirsch , f. r. joaquim and a. vicente , arxiv:1207.6635 [ hep - ph ] .
a. abada , d. das and c. weiland , jhep * 1203 * ( 2012 ) 100 [ arxiv:1111.5836 [ hep - ph ] ] .
j. c. romao , `` the mssm model '' , ` http://porthos.ist.utl.pt/~romao/homepage/ ` ` publications / mssm - model / mssm - model.pdf ` | it has been pointed out that supersymmetric extensions of the standard model can induce significant changes to the theoretical prediction of the ratio @xmath0 , through lepton flavour violating couplings . in this work
we carry out a full computation of all one - loop corrections to the relevant @xmath1 vertex , and discuss the new contributions to @xmath2 arising in the context of different constrained ( minimal supergravity inspired ) models which succeed in accounting for neutrino data , further considering the possibility of accommodating a near future observation of a @xmath3 transition .
we also re - evaluate the prospects for @xmath2 in the framework of unconstrained supersymmetric models . in all cases ,
we address the question of whether it is possible to saturate the current experimental sensitivity on @xmath2 while in agreement with the recent limits on @xmath4-meson decay observables ( in particular br(@xmath5 ) and br(@xmath6 ) ) , as well as br(@xmath7 ) and available collider constraints .
our findings reveal that in view of the recent bounds , and even when enhanced by effective sources of flavour violation in the right - handed @xmath8 sector , constrained supersymmetric ( seesaw ) models typically provide excessively small contributions to @xmath2 .
larger contributions can be found in more general settings , where the charged higgs mass can be effectively lowered , and even further enhanced in the unconstrained mssm .
however , our analysis clearly shows that even in this last case susy contributions to @xmath2 are still unable to saturate the current experimental bounds on this observable , especially due to a strong tension with the @xmath6 bound .
cftp/12 - 004 + pccf ri 12 - 03 + * revisiting the @xmath9 ratio in * * supersymmetric unified models * + * renato m. fonseca@xmath10 , j. c. romo@xmath10 and a. m. teixeira@xmath11 * @xmath10 centro de fsica terica de partculas , cftp , instituto superior tcnico , + universidade tcnica de lisboa , av .
rovisco pais 1 , 1049 - 001 lisboa , portugal @xmath11 laboratoire de physique corpusculaire , cnrs / in2p3 umr 6533 , + campus des czeaux , 24 av
. des landais , f-63171 aubire cedex , france |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.